Full text of "Journal of cancer research."

See other formats

•KR1AL

Columbia 3Hmber*ttp c*fk'
tntfjeCitpof Jleto ^orfe

College of ^fjpsictans anb burgeons

Reference Hibvatp

THE JOURNAL

EDITED BY

William II . Woglom

associate editor
Frederick Prime

editorial committee

The American Association for Cancer Research

Ernest E. Tyzzer

Harvard University

VOLUME VI

Joseph C. Bloodgood

Johns Hopkins University

H. Gideon Wells

University of Chicago

Leo Loeb

Washington University

James Ewixg

Cornell University

BALTIMORE, MD.
1921

CONTENTS

Number 1, January, 1921

Effect of a Reduction of Lymphocytes on the Growth Rate of Transplanted
Spontaneous Tumors in Mice. Frederick Prime 1

A Comparison of the Growth of Mice which Ultimately Develop Carcinoma
with the Growth of Mice which Do Not Develop Carcinoma. T. Brails-
ford Robertson and L. A. Ray 7

Effect of Blood from Immune Animals upon Transplantable Tumors. Isidor
Kross 25

Inoculation of Sarcomatous Tumors into Negro Fowls, with Special Reference
to the Significance of Chromatophores. Yutaka Kon and Tamotsu Fujii. 31

Influence of the Lymphocyte on the Peritoneal Implantation of Sarcoma in
Mice. Ellis Kellert 41

Primary Spontaneous Squamous Cell Carcinomas in Mice. Studies on the
Incidence and Inheritability of Spontaneous Tumors in Mice. Fifteenth
Communication. Maud Slye, Harriet F. Holmes and H. Gideon Wells. . . 57

Proceedings of the American Association for Cancer Research. Thirteenth
Annual Meeting 87

Number 2, April, 1921

A Study of a Lipomyxosarcoma with Comments upon the Origin of the Fat

Cell. Victor C. Jacobson 109

Parabiosis and Tumor Growth. Isidor Kross 121

The Protein Content of the Whole Blood and Plasma in Cancer. Ruth C.

Theis 127

Problems in Cancer Research. Montrose T. Burrows 131

The Influence of Heredity in Determining Tumor Metastases. Studies in the
Incidence and Inheritability of Spontaneous Tumors in Mice. Sixteenth

Report. Maud Slye 139

Proceedings of the American Association for Cancer Research. Fourteenth
Annual Meeting 175

Number 3, July, 1921

Further Investigations on the Origin of Tumors in Mice. VII. Tumor Age
and Tumor Incidence. Leo Loeb 197

Further Investigations of Disturbances of Blood Sugar Equilibrium in Their
Relation to Neoplasia. G. L. Rohdenburg, O. F. Krehbiel, and A.
Bernhard 223

Pregnancy and Tumor Growth. Isidor Kross 245

Is Cancer Mortality Increasing? Wendell M. Strong 251

iii

CONTENTS

Number 4, October, 1921

Cancer and Parasite. Isidor Kross 257

Teratomas and Their Relation to Age. H. E. Himwich 261

An Atypical Adenoma of the Pancreas Originating in Islet Tissue. Harry

Goldblatt 277

The Influence upon the Growth of Transplanted Flexner-Jobling Rat Car-
cinoma of Hydrogen Ions and of Various Salts in Different Concentra-
tions. Kanematsu Sugiura, Helen Miller Noyes, and K. George Falk. . . . 285
Primary Spontaneous Tumors in the Kidney and Adrenal of Mice. Studies
on the Incidence and Inheritability of Spontaneous Tumors in Mice.
Seventeenth Communication. Maud Slye, Harriet F. Holmes, and H.

Gideon Wells 305

A Critical Investigation of the Freund-Kaminer Reaction. Louis Herly 337

Is Cancer a Biological Phenomenon? Some Heretic Thoughts on Cancer.

F. G. Gade 357

The Relation of Muscular Activity to Carcinoma. A Preliminary Report.

Ivar Sivertsen and A. W. Dahlstrom 365

EFFECT OF A REDUCTION OF LYMPHOCYTES ON
THE GROWTH RATE OF TRANSPLANTED
SPONTANEOUS TUMORS IN MICE

FREDERICK PRIME

From Columbia University, George Crocker Special Research Fund, F. C. Wood,

Director

Received for publication April 2, 1920

Recently much interest has been centered upon the hypoth-
esis that the variation in the number of the hemal lympho-
cytes may exert an important immunal function, diminishing
or increasing the resistance of animals to tumor inoculations.
Experimentally an increase or decrease in these cells is most
easily induced by x-rays, and with the improved modern tech-
nique more accurate quantitative estimations of the rays are
obtainable than was possible a few years ago. We owe to
Murphy and Ellis the observation that small doses of x-ray
repeated over consecutive days reduce the lymphocyte count,
without seriously affecting the health of the animal (1). The
experiments hitherto conducted on this phase of the cancer
problem have been largely directed to the study of the effects
of lymphocytic alterations upon transplantable tumors, where
growth had already been well established. Sittenfield (2),
using the Flexner-Jobling rat carcinoma, a. transplantable
tumor which has been observed through many generations,
found that neither increase nor reduction of the lymphoid ele-
ments in the blood had any influence upon either resistance or
susceptibility to tumor growth, an observation which he con-
firmed in a later series of experiments (3). What influence,
if any, might be observed upon the growth of spontaneous
tumors when transplanted into such x-rayed animals has not
as yet been satisfactorily determined. Murphy and Morton
(4) found that such tumors transplanted to x-rayed animals

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 1

FREDERICK PRIME

"acted much as heteroplastic tissues, retrogressing as the lym-
phoid tissue regenerates, " and they also assert that the resist-
ance of animals to the implantation of their own spontaneous
tumor is increased by stimulating the production of lympho-
cytes with x-ray (5). Chambers, Scott, and Russ (6) have
also reported experiments on one primary mouse tumor only.
They observed a lowering of an animal's resistance to implanta-
tion after a single large generalized dose of x-ray. Since, how-
ever, a large proportion of spontaneous tumors either do not
grow when transplanted into other animals or give a very small
percentage of takes, it seemed of practical importance to test
the other aspect of the problem and to ascertain whether resist-
ance to implantation can be decreased by a reduction of the
lymphocytes. If this were possible, it would provide an easy
method of propagating many tumors of value in cancer research,
which otherwise would be lost owing to our inability to trans-
plant them successfully.

The spontaneous tumors used were derived chiefly from a
special strain of "Lathrop" animals, of which this laboratory
had a large number. They were inoculated into mice from the
same "Lathrop" strain, as such a closely related strain offers
a more favorable soil than stock of another breed. Forty-
eight to seventy-two animals in each case were used, twenty-
four being set aside as controls, and twenty-four to forty-
eight kept for x-ray treatment.

To prepare the animals for inoculation they were given
seven consecutive x-ray treatments according to Murphy's
directions for reducing the number of lymphocytes. At the
end of this time a spontaneous tumor was removed from its
host and a small piece, about 0.002 gram in weight, was inocu-
lated into the right groin of both the untreated controls and the
mice whose lymphocytes had been greatly reduced by x-ray.
Before starting the x-ray treatment, as well as twenty-four hours
after the last dose of x-ray had been given, and at weekly inter-
vals thereafter for six weeks, the blood of the x-rayed mice as
well as that of the controls was counted. The blood pictures
all showed a marked fall in the lymphocyte count after the

GROWTH RATE OF TRANSPLANTED TUMORS

administration of seven consecutive doses of x-ray, with a
gradual return to around the normal point by the end of six
weeks. The controls during the early part of the experiment
all showed a marked spontaneous increase in the lymphocyte
count.

The experiments lasted ten weeks or more, no animals being
included in the last series of counts which had not survived for
this period, and many of the series were observed over a period
of twenty weeks. The results were quite different from those
reported by Murphy, who used tumors which had been trans-
planted for many generations. He records that a reduction of
the lymphocytes in the circulating blood was attended by an
increase in the number of takes. In our experiments there was
found to be practically no difference at all in the inoculation
percentage between the untreated controls and those which
had been given a reducing dosage of x-ray. The lymphocyte
count in many of the controls was higher at the end than at
the beginning of the experiment, and there should, therefore,
according to the DaFano-Murphy theory, have been a smaller
number of takes among them; but this was not the case.

Out of a series of 905 animals living at the end of five weeks
which had been treated with x-ray, 72, or 7.9 per cent, showed
tumors, and 92.1 per cent were negative; while out of 740
controls which were living, 6 per cent showed tumors, and 94
per cent were negative. At the end of ten weeks, out of 740
living animals which had been treated with x-ray, 4.3 per cent
showed tumors, and 95.7 per cent showed no growths at all;
and out of 615 controls, 3.7 per cent of the animals showed
tumor growths, and 96.3 per cent were negative. A difference
of 2 per cent in the number of takes at the end of five weeks
between the animals whose lymphocytes had been reduced and
the controls has no significance, and a difference of 0.6 per cent
at the end of ten weeks shows how little evidence there is in
favor of the theory that the lymphocyte plays any important
part in the production of immunity in relation to transplanted
cancer (see chart 1). If the DaFano-Murphy hypothesis that
the lymphocyte plays an important role in the production of

FREDERICK PRIME

immunity were correct, a reduction in the number of tumor
takes should have occurred in these experiments; but, as has
been shown, such was not the case.

That the x-ray treatment was sufficient to reduce materially
the number of lymphocytes in the circulating blood was clearly
demonstrated by the blood counts; and at the end of four weeks,

Chart 1

ample time for some evidence of tumor growth to be seen, the
blood had not yet returned to normal. Whether this condition
applies to spontaneous tumors alone, or to transplantable
tumors in general, will have to be demonstrated for each tumor.

In the series reported, over forty spontaneous tumors were
used and inoculations were made into 2100 mice.

CONCLUSIONS

Decreasing the circulating lymphocytes by small doses of
x-ray does not render mice more susceptible to the inoculation

GROWTH RATE OF TRANSPLANTED TUMORS

of a spontaneous tumor from another mouse of the same
strain.

Mice with spontaneous lymphocytosis are not resistant to
the implantation of a primary tumor from another mouse of
the same strain.

REFERENCES

(1) Murphy, J. B., and Ellis, A. W. M.: Experiments on the role of lymphoid

tissue in the resistance to experimental tuberculosis in mice. J.
Exper. M., 20: 397, 1914.

(2) Sittenfield, M. J. : The significance of the lymphocyte in immunity to can-

cer. J. Cancer Res., 2: 151, 1917.

(3) Sittenfield, M. J.: Further studies on the importance of the lymphocyte

in cancer immunity. J. M. Res., N. S. 33: 463, 1918.

(4) Murphy, J. B., and Morton, J. J.: The lymphocyte in natural and induced

resistance to transplanted cancer. J. Exper. M., 22: 204, 1915.

(5) Murphy, J. B., and Morton, J. J.: The effect of Roentgen-rays on the rate

of growth of spontaneous tumors in mice. J. Exper. M., 22: 800, 1915.

(6) Chambers, H., Scott, G., and Russ, S. : On the action of x-ra3's upon the

transplantation of a spontaneous carcinoma of the rat. J. Path, and
Bacterid., 23: 384, 1919-20.

A COMPARISON OF THE GROWTH OF MICE WHICH
ULTIMATELY DEVELOP CARCINOMA WITH
THE GROWTH OF MICE WHICH DO
NOT DEVELOP CARCINOMA

T. BRAILSFORD ROBERTSON and L. A. RAY

(From the Department of Physiology and Biochemistry, University of Adelaide,
South Australia, and the Department of Biochemistry, University of Toronto)1

Received for publication July 6, 1920
STATEMENT OF THE PROBLEM

Since cancer is a disease of growth, and its incidence in appre-
ciable form is usually deferred until after the attainment of
maturity, a very important factor in our ultimate understanding
of the etiology of cancer must arise out of a knowledge of the
degree to which the preceding development of the animal is
affected by the ultimate aberration of growth. If the factors
which determine the growth and incidence of carcinoma, for
example, are of an essentially accidental character, such as local
irritation plus infection, then the incidence of carcinoma in an
individual will be decided in the main by factors not inherent
in its habit of growth and metabolism, and the preceding
normal development cannot be expected to foreshadow the
impending pathological outcome. If, on the contrary, the
incidence of carcinoma is determined by internal factors of
growth-habit or metabolic peculiarities plus recurrent external
factors, such as local irritation, then the underlying peculiarity
of growth or metabolism may conceivably reveal itself in more
or less marked departures from the average in the preceding
normal development.

1 The expenses of this research were in part defrayed by a grant from the
special medical research fund of the University of Toronto.

8 T. BRAILSFORD ROBERTSON AND L. A. RAY

A comparison throughout the duration of life of the growth
of animals which ultimately develop spontaneous cancer and
those which do not is, therefore, fundamental to our compre-
hension of the etiology of this disease. The fact that such a
comparison has not hitherto been instituted is undoubtedly
attributable to the experimental labor which is essential to obtain
results of any worth, to the expenditure of time which is requisite,
and to the difficulty of an experimental technique which must
obviously be devised to exclude all chance causes of death,
among which must be reckoned epidemic infections. It is
obvious that if epidemic infections were from time to time to
claim a substantial proportion of victims among the experi-
mental animals, it would become impossible to identify the
individuals which, had they survived the infection, would subse-
quently have developed carcinoma.

METHODS OF RESEARCH

For the past six years we have been engaged upon an extensive
series of experiments of which the immediate object was the
determination of the influence of certain glandular extracts or of
normal food constituents in extraordinary excess upon the
growth of otherwise normal white mice. The unusual constit-
uent was added to an otherwise varied and abundant diet of
barley, egg, biscuit, and lettuce. It was essential to our inquiry
that epidemic infections should be excluded, and we were so far
successful that mortality from these diseases, throughout the
duration of their lives, did not amount to more than 4 per cent
of the total deaths in the males, and 1 per cent in the females.
The animals were weighed once a week from the fourth or fifth
week after birth until the thirtieth week, and thereafter once a
fortnight. Individual records of the growth of each mouse were
kept, and its ultimate fate recorded. All lesions manifest to the
naked eye in postmortem examination of the organ were noted,
and sections of formaldehyde-hardened material were prepared
and examined microscopically. We feel confident that nearly
all cases of carcinoma and other new growths were detected and

GROWTH OF MICE AND CARCINOMA

identified except in a very small proportion of cases in which
postmortem examination of the tissues was rendered impossible
by delayed observation of death and consequent decomposition.

The details of our experimental technique and the growth
and mortality statistics thus obtained have been fully described
elsewhere (1).

By segregating, in each experimental group, the records of
those animals which ultimately developed carcinoma, we have
been able to compare the growth of those animals which
developed carcinoma with the growth of those animals which
did not, in eleven different experimental groups, involving a
total of 324 animals, of which 105 developed spontaneous carci-
noma. The primary growth was usually situated in the axilla
or the groin, and metastases occurred in the lungs. In a small
proportion of cases in the other groups and in every case in the
group receiving tethelin throughout the duration of life (tethelin-
fed males) the only new growths detectable by naked eye exami-
nation were situated in the lung; and these, in the tethelin-fed
group, were very small, frequently but 1 to 4 mm. in diameter.
In the more typical cases, involving a primary growth in the
axilla or groin, death usually occurred within one month of the
appearance of a manifest lump.

The diet of the " normal" animals consisted of crushed barley
and water ad libitum, fresh leaves of lettuce twice a week, dry
hard unsweetened biscuit (" pilot bread") once a week, and 5
cc. of mixed white and yolk of egg to every six animals on each
of six days of the week. The egg was employed as the vehicle
for the administration of the various dietary additions to the
other groups; it was generally consumed within a few minutes of
its introduction into the cage.

The pituitary-fed animals received one-half of an anterior lobe
of an ox pituitary per six animals, emulsified in the daily allow-
ance of egg. The lecithin, prepared from yolks of eggs, was
administered in dosage corresponding to 83 mgm. per mouse per
day. The dosage of cholesterol was 42 mgm. per mouse per day.
The dosage of tethelin (a lipin prepared from the anterior lobe
of the pituitary gland) was 4 mgm. per mouse per day. In one

10 T. BRAILSFORD ROBERTSON AND L. A. RAY

group (tethelin-fed males) the administration of tethelin was
continued throughout life. To one group of females (" discon-
tinuous aclministration,,) this dose of tethelin was administered
in three periods of one month each, namely from the end of
the fourth week until the end of the eighth, from the end of the
twenty-first until the end of the twenty-fifth, and from the end
of the forty-second until the end of the forty-sixth week of life.
To another group of females (" brief administration") the admin-
istration was continued only from the end of the fourth until
the end of the twelfth week, and thus ceased before or at about
the attainment of sexual maturity. Weighings, in this latter
group, were discontinued between the thirtieth and the seventieth
week because, mistakenly as events proved, no noteworthy effects
of the administration upon normal growth subsequent to the
thirtieth week were anticipated.

EXPERIMENTAL RESULTS

The incidence of carcinoma in the various experimental classes
and the average duration of life of the animals in each class and
of those which developed carcinoma are enumerated in table l.2
It will be observed that the percentage of incidence of carcinoma
was substantially unaffected by the various dietaries employed,
with the exception that the incidence was extraordinarily low
(17 per cent) in the group of females which received the " brief
administration' 1 of tethelin. Traces of the same effect may be
seen in the low incidence (25 per cent) in the group which received
"discontinuous administration" of tethelin, the normal incidence
in females being, for the strain employed, from 30 to 50 per cent.
We shall see that of all the growth curves herein presented that
for the "brief administration" tethelin group departed most
widely from the type of growth curve generally displayed by
animals which subsequently developed carcinoma. The high
incidence of carcinoma in the pituitary-fed groups may possibly
be fortuitous, or, again, may be attributable to the pituitary

1 Owing to the cost of publication it has been necessary to omit many tables
bearing in detail the data of the experiments. — Editor.

GROWTH OF MICE AND CARCINOMA

tissue, or, yet again, to the fact that the diet of these animals
differed from that of all other classes in containing a proportion
of raw meat.

The results of the comparison of the growth of the animals
which ultimately developed carcinoma with the growth of the
animals which did not are displayed graphically in Figures 1
to 11, in which the full lines represent the growth of the animals
which did not, and the broken lines represent the growth of the

TABLE 1

Incidence of carcinoma in the various experimental classes and the duration of life

of each

AVERAGE

ANIMALS WHICH

AVERAGE DURA-

DURATION OF
LIFE OF THE
ANIMALS WHICH
DEVELOPED
CARCINOMA

CLASS OF ANIMALS

DEVELOPED
CARCINOMA

TION OF LIFE
OF ALL ANIMALS

Males

per cent

days

Normal

767

826

Pituitary

792

823

Lecithin

731

743

Cholesterol

764

848

Tethelin

866

958

Females

Normal

719

721

Pituitary .'

704

689

Lecithin

731

743

Cholesterol

658

590

Tethelin (discontinuous administration) .

800

970

Tethelin (brief administration)

695

677

animals which did develop carcinoma. The comparison is dis-
continued in each case at the date at which half of the carcinoma
animals had died.

An examination of these curves reveals a striking uniformity
of result. In each of the eleven different experimental groups
the carcinomatous animals were superior in weight to the non-
carcinomatous throughout a large proportion of their lives, and
this occurred independently of whether, at the beginning of the
weighings at four or five weeks of age, they were superior, equal,

T. BRAILSFORD ROBERTSON AND L. A. RAY

GROWTH OF MICE AND CARCINOMA

or inferior in weight to the non-car cinomatous animals. The
superior growth occurred during the period of developing sexual
maturity (adolescent growth) and was merely maintained or
slightly diminished or even lost during later life. These facts
are summarized in table 2.

It will be seen that in four out of the eleven groups the animals
which developed carcinoma were initially inferior in weight to

TABLE 2

Relative weights at different ages of animals which developed carcinoma and those
which did not. + indicates superiority of the animals which developed cancer;
— indicates their inferiority in weight, and = indicates substantial equality of
the two groups

70 WEEKS TO

DEATH OF

CLASS OF ANIMALS

INITIAL
WEIGHT (4
TO 5 WEEKS

10 TO 30
WEEKS

30 to 70

WEEKS

ONE HALF OF
ANIMALS
WHICH
DEVELOPED
CARCINOMA

Males.

Normal

Pituitary

Lecithin

Cholesterol

Tethelin

Females

Normal

Pituitary

Lecithin

Cholesterol

Tethelin (discontinuous adminis-

tration)

Tethelin (brief administration)

the animals which did not subsequently develop carcinoma.
Yet even in these cases superiority of weight was ultimately
attained, and in three out of the four groups superiority of weight
was attained before the thirtieth week. In the remaining group
(lecithin-fed females) the initial inferiority was very great
(1.65 grams, or 15 per cent of the mean weight) yet this group
attained equality to the non-carcinomatous animals before the
thirtieth week, and surpassed them thereafter.

T. BRAILSFORD ROBERTSON AND L. A. RAY

— ' J 3 7 * S 7 3 7 Z X 3,

Fig. 3. Comparison of the Growth of Pituitary Fed Male White Mice
Which Ultimately Developed Carcinoma (Broken Line) with
that of Pituitary Fed Male White Mice Which Did
Not Develop Carcinoma (Full Line)

w i 1 — — — i r 1 1 1 1 i i

3 i u h * a m n m m hi

Fig. 4. Comparison of the Growth of Pituitary Fed Female White Mice
Which Ultimately Developed Carcinoma (Broken Line) with
that of Pituitary Fed Female White Mice Which Did
Not Develop Carcinoma (Full Line)

GROWTH OF MICE AND CARCINOMA

INTERPRETATION OF THE RESULTS

We may infer that the animals which ultimately develop
carcinoma foreshadow this outcome in their development during
the earlier portions of the third or adolescent growth-cycle.
They are distinguished from the animals which do not subse-
quently develop carcinoma by their exceptionally energetic growth
during the period of adolescence. Prior to this period they do
not display any uniform characteristics of weight, and subse-
quently to the thirtieth week the growth curve of the carci-

' * * * 3 * 3 7? z z s s sr-

Fig. 5. Comparison of the Growth of Tethelin Fed Male White Mice
Which Ultimately Developed Carcinoma (Broken Line) with
that of Tethelin Fed Male White Mice Which Did
Not Develop Carcinoma (Full Line)

nomatous animals usually remains parallel to the curve for the
non-carcinomatous animals or approaches it, the only marked
exception to this rule being afforded by the group of cholesterol-
fed females in which the animals which subsequently developed
carcinoma display a late accretion of weight between the forty-
fifth and the sixty-fifth week.

The same data have afforded material for another significant
comparison, namely, that of the growth of animals which sur-
vived beyond the average duration of life in any given experi-
mental group with the growth of the animals which failed to
attain the average duration of life (2). In this comparison

T. BRAILSFORD ROBERTSON AND L. A. RAY

it was shown that the long-lived animals are distinguished by
relatively energetic early growth, low variability, steadiness of
weight, and absence of late accretions of weight. The short-
lived animals, on the contrary, display relatively deficient early
growth, high variability, instability of weight, and a marked
tendency to acquire late accretions of weight.

The animals which ultimately develop carcinoma, therefore,
display the general characteristics of the long-lived groups of
animals. At first sight this is not surprising, for the average
duration of life of the animals which develop carcinoma exceeds,

■— T- 1 3 X 3 S 3 J _ I X 5 :

Fig. 6. Comparison of the Growth of Tethelin Fed Female White Mice
(Discontinuous Administration) Which Ultimately Developed
Carcinoma (Broken Line) with that of Tethelin Fed
Female White Mice (Discontinuous Administra-
tration) Wthich Did Not Develop
Carcinoma (Full Line)

in eight of the eleven experimental groups, the average life-
duration of the group. On closer examination, however, this
result is not so readily intelligible as it appears. In the first
place, the carcinomatous animals display the characteristics of
the long-lived animals to an exaggerated degree, and, in the
second place, only a little over two-thirds of the carcinomatous
animals actually belonged to the long-lived group (table 3).
On the other hand, of course, many long-lived animals and, in
fact, most of those which survived for the longest periods, failed
to develop carcinoma at all.

GROWTH OF MICE AND CARCINOMA

We cannot, therefore, conclude that these results are due simply
to the longevity induced by the superior growth-energy of the
carcinomatous animals passively carrying them into the age-
zone in which carcinoma occurs. A majority of all the animals
attain the ages at which one-third of the carcinomatous animals
die. A small minority of animals exceed the age at which the
last carcinomatous animal has died, without developing carci-

TABLE 3

Showing the proportion of animals which developed carcinoma which also lived for
longer than the average duration of life

NUMBER OF

ANIMALS DE-

CLASS OF ANIMALS

TOTAL NUMBER

ANIMALS WHICH

VELOPING CAR-

OF ANIMALS

DEVELOPED

CINOMA WHICH

CARCINOMA

ALSO WERE

LONG LIVED

Males

Pituitary

Lecithin

Cholesterol

Tethelin

Totals

152

Females

Normal

Pituitary

Lecithin

Cholesterol

Tethelin (discontinuous administration) .

Tethelin (brief administration), .

Totals

172

noma at all. Moreover, while the lowest incidence of carcinoma
recorded in table 1 occurs in the shortest-lived group (tethelin,
"brief administration") yet the highest incidence occurs in a
group which lived an average of but twenty-four days longer
(pituitary-fed females), and the next lowest incidence occurs in
the longest-lived group of females (tethelin, "discontinuous
administration' ') .

T. BRAILSFORD ROBERTSON AND L. A. RAY

A remarkable feature of the results is their uniformity. This
becomes the more noteworthy when we consider the variability
of the experimental material. No estimates of variability in
quantitative terms have been computed, because in many cases
the small number of animals which ultimately developed carci-
noma in any one experimental group rendered such a computation
of very little value. Nevertheless, it may readily be perceived
that there is no single type to which the carcinomatous animal

1 1 3 J 3 5 Z 3 1 1

Fig. 7. Comparison of the Growth of Tethelin Fed Female White Mice
(Brief Administration) Which Ultimately Developed Carcinoma
(Broken Line) with that of Tethelin Fed Female White
Mice (Brief Administration) Which Did Not
Develop Carcinoma (Full Line)

adheres. No such simple rule can be formulated as that heavy
animals develop carcinoma. Thus in the tethelin (" brief adminis-
tration ") group one of the animals which developed carcinoma
weighed but 22.5 grams at seventy weeks, or nearly 7.5 grams
less than the average weight at the same age of the non-carci-
nomatous animals of the same experimental group. Similar
phenomena were observed in the other experimental groups.
For this reason, of course, it is impossible to foretell the incidence

GROWTH OF MICE AND CARCINOMA

of carcinoma in any individual animal from its preceding growth.
How then does it arise that the average weight of the animals
which subsequently develop carcinoma is so invariably in excess
of the average weight of non-carcinomatous animals throughout
a large proportion of their lives?

There would appear to be only one possible explanation,
namely that it is superiority of growth relatively to the standard
norm of the strain which foreshadows carcinoma, not energetic
growth per se. The animals employed in this, as in any other

— 1 1 2 1 1 B 7 3 7. 7 &

Fig. 8. Comparison of the Growth of Lecithin Fed Male White Mice
Which Ultimately Developed Carcinoma (Broken Line) with
that of Lecithin Fed Male White Mice Which Did Not
Develop Carcinoma (Full Line)

animal experiment, represented a variety of more or less widely
differing strains, each possessed of its own standard of absolute
and relative development. In any sufficiently large group of
animals chosen at random from the general stock, each of the
main strains would be represented in proportion to its frequency.
Since the carcinomatous animals, although so variable among
themselves, invariably displayed an average weight which for a
large period of their lives exceeded the average weight of the
remainder, we must conclude that each of the various strains

T. BRAILSFORD ROBERTSON AND L. A. RAY

in the stock furnished its proportionate quota to the carcinoma-
tous group, and in each strain the animal's which ultimately
developed carcinoma were superior in energy of growth to the
standard norm of the strain. This fact precludes the idea that
we may be dealing with linked size and carcinoma-inheritance,
the carcinomatous animals representing a single strain or a small
number of strains distinguished also by large size. The two
factors, that of absolute size and that of tendency to develop
carcinoma, are, on the contrary, separable.

Man

^ 2 3 J 3 7s * 3 21 * 1.

Fig. 9. Comparison of the Growth of Lecithin Fed Female White Mice
Which Ultimately Developed Carcinoma (Broken Line) with
that of Lecithin Fed Female White Mice Which Did Not
Develop Carcinoma (Full Line)

In the article dealing with the comparison of the growth of
long-lived with that of short-lived animals, it was sought to
interpret the results in terms of the competition between paren-
chymatous and connective tissues during the development of
the animals. In those animals in which the anabolism of paren-
chymatous tissues is exceptionally rapid, the growth of paren-
chymatous structures will be hastened and facilitated. This
may or may not tend to increased weight of the animal at the
time, according to the degree of acceleration experienced and the

GROWTH OF MICE AND CARCINOMA

relative masses of parenchyma and connective tissues in the
body of the animal. It will, however, manifestly delay, through
previous appropriation of foodstuffs, the senescent accretion of
connective tissues, and consequently increase the duration of
life.

If we apply this hypothesis to the results enumerated in this
paper we must conclude that the animals which subsequently
develop carcinoma are distinguished by exceptionally rapid
anabolism and growth of parenchyma which, during adolescence,
more than compensates for the complementary delay in con-

CMLOTCIIM.

1 I Z 35 * "3 3 3 J 2 £ S J

Fig. 10. Comparison of the Growth of Cholesterol Fed Male White
Mice Which Ultimately Developed Carcinoma (Broken Line)

WITH THAT OF CHOLESTEROL FED MaLE WHITE MlCE WHICH

Did Not Develop Carcinoma (Full Line)

nective tissue growth. In the absence of carcinoma this would
result in great longevity, which however, is curtailed by a patho-
logical proliferation of parenchymatous tissue. We may suppose
that in such animals the energy of response to some external
stimulus, such as a local source of irritation, is so excessive that
when the age of the animal is sufficiently advanced to have
resulted in some weakening of the competition of other paren-
chyma (through the parasitism of connective tissues) the effect
is to institute a growth which appropriates nutriment to itself
at the expense of all the adjacent tissues.

T. BRAILSFORD ROBERTSON AND L. A. RAY

In accordance with this hypothesis we find that the incidence
of carcinoma is lowest in that experimental group in which the
growth curve departs most widely from the long-lived type,
namely, in the tethelin " brief administration" group in which
there is an extraordinary late accretion of tissue (3). The only
other group of carcinomatous animals which reveals late accre-
tion of weight is also the group which has the briefest duration
of life (cholesterol-fed females, table 1).

9 13 29 19 49 9| «9 T9 tS M

Fig. 11. Comparison of the Growth of Cholesterol Fed Female White
Mice Which Ultimately Developed Carcinoma (Broken Line)
with that of cholesterol fed female white mlce which
Did Not Develop Carcinoma (Full Line)

The longest-lived animals in any group must, therefore, be
those in which the speed of anabolism and energy of growth of
the parenchyma just fall short of the tendency to respond abnor-
mally, by the formation of neoplasms, to recurring local irritative
stimuli.

The comparative growth of animals which ultimately devel-
oped neoplasms other than carcinoma will form the subject of
subsequent communications.

GROWTH OF MICE AND CARCINOMA

SUMMARY

1. The incidence of carcinoma in mice is foreshadowed in
their preceding development; but, owing to the variability of
the animals, it is not possible to foretell the incidence of carci-
noma in any single animal from its curve of growth.

2. The animals which ultimately develop carcinoma are
distinguished by relatively energetic growth during the period
of adolescence. The lead over the other animals which is thus
established is usually maintained throughout life, but is some-
times lost in the later periods of life through late accretion of
weight by the animals which do not develop carcinoma.

3. These results do not admit of interpretation by supposing
that superior growth favors longevity, and thus passively carries
the animals into the age-zone in which carcinoma occurs. A
majority of all the animals surpass at death the ages at which
one-third of the carcinomatous animals die. There is no cor-
relation between the percentage incidence of cancer in any group
and its average duration of life; but animals displaying an aver-
age growth curve which departs widely from the carcinomatous
type also display a low percentage incidence of carcinoma.

4. The results are not due to linked inheritance of size and
tendency to develop carcinoma, since the factors of absolute
size and development of carcinoma are separable.

5. The results are interpreted to mean that the animals which
ultimately develop carcinoma are those in which the anabolism
and, therefore, the growth of parenchyma is exceptionally rapid.
In such animals the energy of response to local irritative stimuli
may be so excessive as to overcome the competition of other
tissues and initiate a new growth.

6. The longest-lived animals in any group are, therefore, those
in which the speed of anabolism and energy of growth of the
parenchyma just fall short of those which lead, under the influ-
ence of recurrent irritative stimuli, to the formation of new
growths.

REFERENCES

(1) Robertson, T. B., and Ray, L. A.: Jour. Biol. Chem., 1916, xxiv, 347, 363,

385, 397; 1916, xxv, 635, 647; 1919, xxxvii, 377, 393, 427, 443, 455.

(2) Robertson, T. B., and Ray, L. A.: Jour. Biol. Chem., 1920, xlii, 71.

(3) Robertson, T. B., and Ray, L. A.: Loc. cit.

EFFECT OF BLOOD FROM IMMUNE ANIMALS UPON
TRANSPLANTABLE TUMORS

ISIDOR KROSS

From Columbia University, George Crocker Special Research Fund, F. C. Wood,

Director

Received for publication October 22, 1920

Even a cursory review of research in experimental cancer will
reveal many inconsistencies, various workers having obtained
contradictory results when attacking a problem with apparently
similar methods. In many cases, however, patient repetition
with careful checking up of the findings has brought to light
factors which offer a satisfactory explanation for these puzzling
variations. In this study an attempt was made to clear up one
of the problems in which different workers have reached dia-
metrically opposed conclusions.

Jensen (1) reported the disappearance of transplanted tumors
in mice injected with serum from rabbits treated with carcinoma.
In a later publication (2) he stated that while he realized that
spontaneous retrogression would explain most of his "cures,"
it did not account for the disappearance of several of the very
large tumors.

von Leyden and Blumenthal (3) reported the successful out-
come of the treatment of a dog with a carcinoma, the diagnosis
of the growth having been verified by microscopical examination.
The method was as follows: Several rabbits were given a series
of subcutaneous injections of carcinoma emulsion over a period
of several weeks, and the serum obtained from these rabbits
was injected subcutaneously into the dog.

Clowes (4) reported the results of experiments with mice in
which one half of the animals received injections of "immune"
serum obtained from mice whose tumors had retrogressed spon-
taneously. The other half acted as controls and received normal

ISIDOR KROSS

mouse serum. Of the twenty mice in the first series, only one
failed to show the beneficial effect of the immune serum, and all
were alive at the time of the report; while in the control series,
five were dead and all the others had tumors exceeding in size
those of the first series. Clowes and Baeslack (5) reported the
results of experiments in which one lot of mice was treated with
tumor mixed with immune serum, a second with tumor mixed
with normal serum, and a third with tumor mixed with normal
saline solution. The difference between the results obtained in
the two latter groups was negligible. The difference between
the results in the "immune" set and in the two control sets,
however, was considerable, 31.6 per cent of tumors developing
in the controls and only 12.3 per cent in the immune group. On
the basis of these findings, they justified their assumption of the
presence in the blood of immune bodies antagonistic to the devel-
opment of carcinoma.

Crile and Beebe (6) reported a series of ten blood transfusions
in dogs with an infectious lymphosarcoma, resulting in cure
in seven cases and in marked improvement in two. In the
remaining case, in which the treatment was a complete failure,
it was later found that the donor was not an immune animal.

von Dungern (7) obtained serum from rabbits in which there
has occurred spontaneous retrogression of a sarcoma. Seven
rabbits were treated by intravenous injection of this serum before
being inoculated with sarcoma. Nine other rabbits were em-
ployed as controls. Six of the control animals developed tumors;
in none of the treated animals did a tumor develop. Uhlenhuth,
Handel, and Steffenhagen (8), however, performed the same
experiment and found that rats treated with "immune" serum
gave 94 per cent of successful takes, whereas those treated with
normal serum gave 83 per cent, and those that received no treat-
ment at all gave only 66 per cent. In another series of experi-
ments, these authors mixed an emulsion of tumor tissue with
immune serum. The injection of this mixture gave 100 per
cent successful takes. In their experience heterologous immune
serum proved just as unsuccessful as the homologous serum.

EFFECT OF BLOOD UPON TRANSPLANTABLE TUMORS 27

Haaland (9) treated Berlin mice with the serum of immune
Hamburg mice. On subsequent inoculation of the Berlin mice
with a sarcoma to which they were normally susceptible the
tumors grew just as well as they did in the untreated controls.
He then repeated the experiments, employing the serum from
immune Danish mice; but here again there was no difference
between the growth of the tumors in treated and in untreated
animals.

Russell (10) and Bashford, Murray, and Cramer (11) also
failed to obtain positive results with the blood from immune
animals.

Sisto (12) repeated the experiments of von Dungern (13) and
others in attempting to influence tumor growth by the injection
of spleno toxic, orchidotoxic, and hepato toxic sera from rabbits
into which emulsions of the fresh respective rabbit organ had
been injected. His published charts show practically no effect
of the injections one way or the other.

It would not be amiss in a paper of this character to make
mention of a closely allied but not quite similar method of treat-
ment that has been tried in man. In 1910 Hodenpyl (14)
published a preliminary report of the treatment of carcinoma
by the subcutaneous and intravenous injection of ascitic fluid
from a patient who, he believed, had undergone a spontaneous
recovery from carcinoma. All of his patients showed subjective
improvement. It is, however, well known to those who have
had experience in the treatment of inoperable carcinoma that
every new form of therapy along this line is followed by a
temporary subjective improvement, i.e., relief of pain, diminution
of weakness, improvement in appetite, etc., only to be succeeded
by the inevitably fatal termination. Hodenpyl stated that the
tumors diminished in size and that some disappeared. No cures
were obtained, however, and the donor of the ascitic fluid also
died later from cancer.

Ill and Miningham (15) repeated Hodenpyl's treatment in a
series of twenty-seven cases in which ascitic fluid was injected
subcutaneously. The fluid was obtained from a patient who
was apparently recovering from carcinoma of the liver; but

ISIDOR KROSS

autopsy later on revealed widespread carcinoma of the liver;
ovaries, and intestines. The authors noted marked subjective
improvement, although they too failed to produce a single cure
with this treatment.

In the experiments about to be reported, the procedure was
as follows: The rats used were those known in the laboratory
as the Marshall and the August breeds. The Marshall rats are
susceptible to the Flexner rat carcinoma and refractory to the
Jensen rat sarcoma. The August rats have the reverse char-
acteristics in respect to these two tumors. To make doubly
sure that the animals used were immune, a preliminary inocu-
lation of the Flexner rat carcinoma (FRC) was made in the
August, and of the Jensen rat sarcoma (JRS) in the Marshall
rats. Only those animals were used which did not develop
tumors or in which there was a small growth with subsequent
complete retrogression. These animals constituted the stock
from which the immune blood was taken.

One group of Marshall rats was treated by intraperitoneal
injections of blood from the August rats which had showed a
natural immunity against the Flexner rat carcinoma. A second
group of Marshall rats received an intraperitoneal injection of
blood from normal Marshall rats. A thud group received no
preliminary injection of blood. All three groups were then
inoculated with grafts of 0.002 gram of the FRC in the right
axilla. Exactly the same plan was followed with the August
rats, the immune Marshall rats being used as the source for the
immune blood, and the JRS being inoculated. The amount of
blood injected varied from 2 to 3 cc. depending upon the size
of the animal, the idea being to take as much blood as possible
without killing the rat. The tumor inoculation was made
synchronously with the blood injection or within a period of
forty-eight hours after it.

In accordance with the theory of the workers cited, there should
have been a complete absence of growth or at least a noticeable
retardation of growth in those animals that had been treated
with blood taken from the immune animals. Examination of the
records, however, showed that the rats treated with the immune

EFFECT OF BLOOD UPON TRANSPLANTABLE TUMORS 29

blood not only did not show any absence or retardation of growth
of the tumor, but actually showed tumors that were in most
instances larger, and developed earlier than those in the un-
treated animals. Even the animals treated with normal blood
showed this phenomenon though in a lesser degree. Gay (16)
found that injection of blood from insusceptible or refractory
animals lead to an increase in the number of takes of carcinoma
implantation.

Another series of experiments was made in order to determine
whether the blood or other proteins injected act as a food. Three
sets of rats, 36 in each group, were employed. One group
received 1 cc. of blood (18 rats subcutaneously, and 18 intra-
peritoneally) every three days over a period of six weeks. A
second group received 1 cc. of egg-white (18 subcutaneously,
and 18 intraperitoneally) every three days over a similar period.
A third group was kept as controls. All these animals were
inoculated with the JRS, and the tumors were charted every week
for eight weeks. No appreciable difference was noted in the
growth of the tumors in the three groups. Hence, the possi-
bility that the injections acted as nutriment could be disregarded.

CONCLUSIONS

It is apparent from these experiments that :

1. If immune bodies do exist in animals that are refractory
to tumor growth, they are not resident in the circulating blood.

2. The transfusion of blood, if it has any influence, accelerates
the development of a tumor, as regards both time and intensity
of growth; the inadvisability of transfusing human cancer cases
is therefore evident.

3. The results obtained by those investigators who report
successful cures after injections of blood or serum must be
explained by the assumption that they were dealing either with
infectious granulomata or with tumors that disappeared
spontaneously.

In order to reduce the effects of extraneous factors as much as
possible, the tests were conducted in eight successive series, each

ISIDOR KROSS

set containing five test animals and ten controls. The test'
animals were those treated with preliminary injections of "im-
mune" blood, the controls those that received injections of
normal blood or nothing at all.

REFERENCES

(1) Jensen: Centralbl. f. Bakteriol., 1 Abt., Orig., 1903, xxxiv, 30.

(2) Jensen: Ztschr. f. Krebsforsch., 1909, vii, 281.

(3) vox Leyden and Blumenthal : Deutsch. med. Wchnschr., 1902, xxviii, 637.

(4) Clowes: Bull. Johns Hopkins Hosp., 1905, xvi, 130.

(5) Clowes and Baeslack: Med. News, 1905, lxxxvii, 969.

(6) Chile and Beebe: J. M. Res., 1908, xiii, 385.

(7) v. Dungern: Ztschr. f. Immunitatsforsch., 1910, v, 695.

(8) Handel and Steffenhagen: Arb. a. d. k. Gsndhtsamte., 1911, xxxvi, 490.

(9) Haaland: Berl. klin. Wchnschr., 1907, xliv, 717.

(10) Russell: Third Sci. Report, Imperial Cancer Research Fund, London,

1908, p. 341.

(11) Bashford, Murray, and Cramer: Third Sci. Report, Imperial Cancer

Research Fund, London, 1908, p. 315.

(12) Sisto: Tumori, 1918, iv, 32.

(13) v. Dungern: Die Antikorper, Jena, 1903, p. 30.

(14) Hodenpyl: Med. Rec, 1910, lxxvii, 359.

(15) III and Miningham: J. A. M. A., 1912, lix, 497.

(16) Gay: J. M. Res., 1919 xv, 186.

INOCULATION OF SARCOMATOUS TUMORS INTO
NEGRO FOWLS, WITH SPECIAL REFERENCE TO
THE SIGNIFICANCE OF CHROMATOPHORES

YUTAKA KON and TAMOTSU FUJII
Tokio

Received for publication December 22, 1920

A kind of domestic fowl, called in western countries "silky
fowl" or "negro fowl," has its origin in the Orient. Their most
striking characteristics are, first, the extra toes on the inner
side of the foot, showing a complete or incomplete adherence of
first and second of them, and second, marked pigmentation of
every part of the body. Although the feathers of genuine negro
fowls are pure white or pure black, the skin itself always appears
extremely bluish brown in color and contains a number of chro-
matophores in the cutaneous and subcutaneous connective tissue.
The periosteum and endosteum, dura and pia mater, perivascular
connective tissue and serosa also show the brown coloration,
diffuse and spotted. Of the organs the testicles and ovaries
are most strongly tinted; the lungs show spotted pigmentation
as in anthracosis. The parenchymatous organs, namely, the
liver, kidney, and spleen, are usually free from abnormal pig-
mentation. The coloration of bone is so remarkable that we
suspected at first its identity with animal ochronosis.

Microscopically we ascertain that the brown coloration every-
where in the body is due to the existence of a large number of
chromatophores in the interstitial tissue, especially along the
course of the blood-vessels. The chromatophores assume a
form quite similar to those in normal fowl, having a number of
irregular long and slender processes and fine yellowish brown
pigment granules with an entire similarity to melanin in micro-
scopical appearance. Pigment granules in the cells which are in

YUTAKA KON AND TAMOTSU FUJII

resting stage are very fine and equal in size. When stimulated
by experimental means they become irregular in size, and the
cells lose their processes, taking a round shape like any other
plastic cells.

The brown coloration of the bone is caused by the pigment
cells in the periosteum. The endosteum of the wing bone, which
covers the inner surface of empty medullary cavity, has also a
number of chromatophores. The matrix of the bone is free
from any kind of pigment, in contradistinction to the condition
found in animal ochronosis, while the bone-corpuscles contain
genuine melanin granules in the protoplasm, which cause a slight
brown coloration of the bone tissue. This is especially evident
in internal and external layers, near the peri- and endosteum.
The bone marrow also contains chromatophores in greater or
less number.

The cartilage shows generally no marked coloration. The
cartilage cells, however, sometimes contain microscopically a
number of pigment granules. The dura mater is deep brown
like the periosteum. The pia mater shows grayish brown spots
due to pigment cells. The fascia is more or less brown in color,
owing to a number of chromatophores which are sometimes
to be seen in intermuscular connective tissue. The periosteum,
pleura, and synovial membrane also show spotted pigmentation.
The pigmentation of the testicles and ovaries is caused by the
existence of chromatophores in the interstitial tissue, while the
tubules and the follicles have no pigment particles. A few
pigment cells are found in the capsules of the spleen, liver,
kidney, and pancreas, and in the perivascular connective tissue.
The parenchyma of the organs and the stellate cells in the liver
contain no pigment granules.

The principal subject of our investigation with this kind of
fowl was to determine whether or not the chromatophores take
part in the neoplastic proliferation brought about by the experi-
mental inoculation of transplantable tumors. Such participa-
tion, if found to occur, would be undoubtedly reliable evidence
of the secondary acquirement of the neoplastic properties by
normal tissue. Another object was to ascertain the origin of

INOCULATION OF SARCOMATOUS TUMORS

melanotic pigment in bone and cartilage cells in negro fowls,
that is, whether it is formed in osteoblastic and chondroblastic
cells themselves or is given to them by adjacent chromatophores.
Moreover, we expected this material to be most suitable for the
purpose of studying the relation between the pigment in epi-
thelial cells and chromatophores, which has been an interesting
subject of discussion among authors.

Fig. 1. Beginning myxosarcomatous growth in the endosteum of the wing
bone of a negro fowl, showing a number of chromatophores in the periosteum and
endosteum. The chromatophores in the endosteum are deranged and replaced
by the new grown tissue. Bone corpuscles contain melanin granules, especially
evident near the periosteum.

I. THE INOCULATION OF TRANSPLANTABLE TUMORS INTO NEGRO

FOWLS

Two different strains of transplantable tumors (chondro-
sarcoma and myxosarcoma) were chosen for this purpose. Fifty
negro fowls were inoculated, 15 of them with chondrosarcoma
and 30 with myxosarcoma. We usually utilized for inoculation
the empty medullary cavity of the wing bones, because the

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 1

YUTAKA KON AND TAMOTSU FUJII

endosteum is rich in chromatophores and if any proliferation
should take place it could be easily followed. The wing bone
was exposed, perforated by means of a borer, and inoculated
with the tumor. In the case of chondrosarcoma a small piece
of tissue was used; in that of myxosarcoma, except in a few
instances, 2 to 4 per cent watery filtrate of the tumor extract.
The tumors were easily transplantable in the medullary cavity
of the wing bone. Positive transplantations were obtained in

- », •:

' . ,<\ „v, .a. -.

w£l ' - i - - •> - '

* :* % - . ' - *

^^^^^^^^^^^^^ . .'--'^ .C-

Fig. 2. Advanced myxosarcomatous tumor of a negro fowl, showing a num-
ber of melanotic tumor cells.

9 out of 15 inoculations with chondrosarcoma and nearly in all
with myxosarcoma.

The development of chondrosarcoma results in a very great
reactive formation of callus which consists of connective, carti-
lage, osteoid, and bone tissues. The neoplastic tissue, however,
differs from the reactive callus tissue which is accompanied by
a number of chromatophores. The chondromatous cells do not
contain any particles of pigment. Briefly, in the series of inocu-
lations of chondroma no neoplastic change of ordinary chro-
matophores was seen. On th(4 contrary, the myxomatous tumor

INOCULATION OF SARCOMATOUS TUMORS

fills up the medullary cavity of the wing bone, accompanied by
some reactive callus tissue. In our cases metastasis was fre-
quently to be seen, especially multiple, into lungs and medi-
astinum, exceptionally into the skin, peritoneum, pericardium,
and liver. The neoplastic tissue in the medullary cavities of
wing bones is somewhat gray in color. Some of the metastatic
tumors also show a grayish coloration.

Fig. 3. The same specimen in higher enlargement.

Microscopically the neoplastic proliferation of endosteum in
cases of inoculation with filtrate of tumor extract seemed to
begin in the connective tissue in the layer close to the bone
surface. The chromatophores in the endosteum are deranged
and replaced by the new grown tissue. In this stage they did
not show any sign of proliferation, having the usual form with
a number of processes. Sooner or later, however, they began
to multiply and produced numerous round cells with coarse
pigment granules of irregular size. They were prolonged and
transformed into long spindle cells which assumed an appear-
ance quite similar to genuine tumor cells, not only in their form,
but in their arrangement in the tumor tissue. Pigment granules

THE JOURNAL OF CAXCER RESEARCH, VOt. VI, XO. 1

YUTAKA KON AND TAMOTSU FUJII

are generally found close to the nucleus. Most of them, there-
fore, are to be considered as true neoplastic elements. The
pigment may not be a general metabolic product of negro fowl,
for the granules are found most abundantly near periosteum and
endosteum, where chromatophores exist in large number. This
rinding confirms that of Rous, Fujinami, Hayashi and others,
that the normal tissue may acquire neoplastic nature by the

Fig. 4. Advanced myxosarcomatous tumor of a negro fowl with an enormous
proliferation of melanotic cells.

inoculation of the tumor tissue or its extract. The formation
of a pure melanotic tumor is not to be expected, for the pigment
cells are always accompanied by a large number of the usual
connective tissue cells, which are more readily proliferative
than the former. In further generations of inoculation of tumor
tissue with melanotic elements into the normal fowls the pig-
mented tumor cells gradually disappear and the neoplastic
proliferation of neighboring tissue eells overcomes the trans-
planted pigmented tumor cells.

INOCULATION OF SARCOMATOUS TUMORS

II. THE RELATION BETWEEN THE MELANOTIC PIGMENT IN EPI-
THELIAL CELLS AND CHROMATOPHORES

With regard to the origin of the melanotic pigment in chro-
matophores and in epithelium a difference of opinion exists.
Some authors maintain that the pigment originates in chro-
matophores and then goes over into epithelium, while others
assert that it is primarily produced in epithelial cells. At the
present time a majority of authors seem to consider that mela-
notic pigment is formed independently both in epithelium and
in chromatophores.

The negro fowl offers a most suitable characteristic for the
purpose of solving this problem. As mentioned, there are two
kinds of negro fowl, white feathered and black feathered. Mela-
nin granules are not to be found in the epithelium of the white
feathered, while the black feathered variety has a number of
pigment granules in the epithelial cells of the skin, more abun-
dantly in the exposed parts such as the legs or the crest. In
other parts of the body, where stratum corneum is thick, the
melanin content of epithelium is generally light. The chro-
matophores in cutaneous and subcutaneus connective tissues
differ in numbers among both kinds of negro fowls. In this point
the white feathered negro fowl is to be compared with the ape
(Cercopithecus mono), which has white hair and no pigment in
the skin, but a large number of chromatophores in the cutaneous
connective tissue. On the other hand the embryonal skin of
the normal black feathered fowl shows the pigment only in the
epithelial layer in the early stage. We recognized, therefore,
that melanin pigment may be formed in both elements inde-
pendently. The migration of the pigment granules from one to
another seems to be disproved in this case, although the general
idea that the pigment of the eye is formed in the retina and
secondarily given to the cells of the choroid, may not be
impossible.

YUTAKA KON AND TAMOTSU FUJII

III. THE ORIGIN OF MELANOTIC PIGMENT IN BONE CORPUSCLES -
AND IN CARTILAGE CELLS

It is rather important to determine whether the above men-
tioned pigment in bone and cartilage cells is formed in the cells
themselves or is given from the adjacent chromatophores. For
this purpose I have intentionally broken the ulna of negro fowls
in one or both sides and examined the callus histologically in the
course of one to forty-three days after fracture. Twelve negro
fowls were used for this experiment. During healing of fractured
bones the callus formation takes place at first close to the bone
tissue, stripping off the layer of chromatophores. Periosteal
chromatophores are arranged close to each other in the earlier
stage. With the advance of callus formation they are separated
by new grown tissue cells. Actual proliferation of chromato-
phores takes place in some degree, forming round shaped cells
with irregular melanin granules. At the same time melanin
granules are to be found in some typical osteoblasts and chondro-
blasts in the callus, but far less in number than in new formed
chromatophores. Pigment granules in the plastic cells are
generally coarse in size, but they do not show any microchemical
difference from those in chromatophores. The amount of pig-
mented plastic cells is generally proportional to the number of
chromatophores in the body. There is no actual sign of trans-
ference of pigment from chromatophores to plastic cells. The
metaplastic change of chromatophores to fixed tissue cells is
not recognizable, for the latter are to be acknowledged undoubt-
edly as a sort of wandering cell. We conclude, therefore, that
genuine osteoblasts and chondroblasts themselves can make
melanin into protoplasm, and that the melanin formation may
occur in many kinds of tissue cells, if there are some melanogenetic
compounds from which melanin could be derived, and fermenta-
tive (perhaps oxidative) elements which act on them.

SUMMARY

1. Oriental negro fowls have a Large number of chromatophores
in all parts of the body, especially in the connective tissue, dura

INOCULATION OF SARCOMATOUS TUMORS

and pia mater, periosteum and endosteum, serous membrane,
interstitial tissue of genital glands, and in the lungs.

2. The bone corpuscles and occasionally the cartilage cells
also contain melanotic pigment in their protoplasm.

3. Myxosarcomatous tumors of negro fowls grown by inocu-
lation of tissue or its extract assume a melanotic nature, probably
due to the neoplastic proliferation of chromatophores. This
is actual evidence for the acquisition by normal tissue of neo-
plastic nature.

4. The melanotic pigment in the epithelium and connective
tissue is formed in both elements independently of each other.
The migration of pigment granules from one to another is not
recognizable.

5. Osteoblasts and chondroblasts in callus also form melanotic
pigment in their protoplasm a.

INFLUENCE OF THE LYMPHOCYTE ON THE PERI-
TONEAL IMPLANTATION OF SARCOMA IN MICE

ELLIS KELLERT
From the Bender Hygienic Laboratory, Albany, New York
Received for publication January 22, 1921

No phase of tumor investigation in recent years has aroused
greater interest than that of the relation of the lymphocyte to
natural and induced resistance to cancer. Following the obser-
vation by Da Fano (1) that lymphoid and plasma cells are
constantly associated with degenerating tumors and that the
lymphocyte might be the agent which distributes immunity
throughout the organism, many articles have been written
tending to prove an antagonistic action between the lymphocyte
and tumor growth. Burgess (2) noted that the retrogressing
implant of cancer in mice is invariably surrounded by large
numbers of lymphocytes. Similar observations were made by
v. Dungern and Coca (3) . Apolant (4) found increased difficulty
in immunizing splenectomized animals to tumor implants.
Tyzzer (5) working with a Japanese waltzing-mouse tumor found
that implants grew vigorously for a time in non-susceptible
mice, but were destroyed soon after by an inflammatory reaction
excited only by living tumor; the cell infiltration disappearing
as the tumor became necrotic. In previously immunized non-
susceptible mice the process of destruction was the same but
began much earlier. Braunstein (6) concluded that the spleen
possesses highly developed powers of resistance against tumor
growth, while splenectomized animals are more liable than normal
animals to inroads of malignant growth. Rohdenburg and
Johnston (7) made parallel observations upon thymus, pancreas,
spleen, hypophysis, and testis, showing that after extirpation of
the thyroid, thymus, or testes, the animal had a lessened resist-
ance to carcinoma.

ELLIS KELLERT

Baeslack (8) studied the blood of inoculated animals and
reported a decrease in the numbers of circulating lymphocytes
in mice with growing tumors. Murphy (9), working with tissue
cultures, noted that resistance to heteroplastic tissue grafts
apparently depends upon the activity of the lymphocyte and
obtained similar results with experimental cancer in mice.
Murphy and Morton (10) in a series of experiments with trans-
planted and spontaneous tumors in mice noted that slight expo-
sure to Roentgen rays, inducing a lymphocytosis, brought about
a more highly resistant state to implanted tumor; whereas larger
doses of the z-ray, inducing a leucopenia, caused a decreased
resistance to inoculation with tumor, and that animals naturally
immune were rendered susceptible by this method. They con-
cluded that destruction of the lymphocytes by the z-ray causes
a loss of the natural or induced resistance to the growth of inocu-
lated cancer.

The apparently conclusive results of the investigators referred
to above have not been without contradiction. Wade (11)
noted that with the establishment and growth of an inoculable
tumor, there is a steady increase in the percentage of lympho-
cytes in the circulating blood. Rous and Murphy (12), experi-
menting with sarcoma of the fowl, decided that no fundamental
importance can be attributed to the protective function of
lymphocytes in mammals and birds alike although they did
state that the " lymphocyte has an association with the process
of resistance in the fowl more marked than that seen in mammals."

Serafini (13) observed that ligation of the splenic vessels in
rats and the injection of splenic tissue does not promote the
growth of implanted tumor. Jones and Rous (14) noted that
mice inoculated eight to ten days after splenectomy have the
same resistance as normal animals; and that mice inoculated
three and a half weeks after splenectomy live longer than normal
animals. Bullock (15) and Stevenson (16) could not find that
the presence of spleen inhibited the growth of tumor implanted
into the chick embryo. Sittenfield (17), having utilized various
means of inducing a hyperlymphocytosis, stated that "neither
increase nor reduction of the lymphoid elements in the blood

PERITONEAL IMPLANTATION OF SARCOMA

had any influence upon either resistance or susceptibility to
tumor growth." Stevenson (18), again, found in 1917 that
tumors show no inhibition of growth when grown on the allantois
of the chick embryo for ten days in the presence of chicken
spleen. Bullock and Rohdenburg (19) (20) also reported that
removal of the spleen does not favor the growth of heteroplastic
tumor grafts. More recently Prime (21) found that, following
lymphocytic reduction by exposure to the x-ray, no appreciable
decrease in immunity to tumor inoculation occurs.

As a result of these conflicting and widely divergent results
we find that the role of the lymphocyte in immunity to cancer
is still undetermined. The varying conclusions arrived at are
due to the methods employed and to the instability of the
leucocyte count in the lower mammals, for blood-cell counts in
mice may be extremely variable. In attempting to devise a
new means of approaching the problem, it occurred to the author
that intraperitoneal inoculation of a tumor, preceded and fol-
lowed by study of the cellular elements of the peritoneum, might
yield information regarding the part played by the lymphocyte
in cancer immunity.

The normal peritoneum of the white mouse contains numerous
cellular elements, among which are found in predominating
numbers the large mononuclear phagocytic cells designated as
macrophages and small mononuclear cells having the morphology
and staining reactions of lymphocytes. In addition there are
large cells whose cytoplasm is filled with coarse basophilegranules,
resembling the mast cells of the blood, a few polynuclear leuco-
cytes, and erythrocytes rarely. The macrophages, as nearly
as can be determined by direct counts, are found to vary from
10,000 to 20,000 per cubic millimeter, whereas the lymphocytes
are far more numerous, usually about 115,000 per cubic milli-
meter. In a few mice the macrophages predominate, but in
the majority the lymphoid cells are more numerous. In differ-
ential counts made by obtaining peritoneal fluid with capillary
glass pipettes, the average percentages found in a series of 24
mice were as follows :

ELLIS KELLERT

per cent

Macrophages

Large mononuclears

Lymphocytes

Poymorphonuclear neutrophiles.
Basophi es

34.3
7.0

55.0
3.5
0.2

100.0

The large mononuclears in the above classification are much
larger than the lymphoid cells, but do not possess the abundant
foamy cytoplasm of the macrophage. The designation of the
small mononuclear cells as lymphocytes may be questioned.
Evans (22) apparently refers to certain small mononuclears in
the peritoneum as young macrophages, asserting that they will
take up granules of vital stains, a characteristic lacking in true
lymphocytes. This may apply to the cells referred to above as
large mononuclears. Morphologically we find the small cells
to be identical with the lymphoid cells of the blood. We do
not find that they possess phagocytic properties for bacteria,
inert particles, or vital stains; and finally, they decrease markedly
under z-ray exposures.

It is very evident, then, that the peritoneum of the mouse
contains large numbers of cells having all the characters of
lymphocytes, and it is the purpose of the present investigation
to show their influence on intraperitoneal inoculations of a
mouse tumor obtained from the Crocker Research Fund, where
it is designated as no. 180, and is regarded as a sarcoma. The
tumor grows readily, and in our hands has not failed in more than
5 per cent of subcutaneous and 2 per cent of intraperitoneal
inoculations. Inoculations were made by means of a small
trocar needle.

The white mice selected were young adult animals; an equal
number of each sex. Smears from the peritoneum were made
before inoculation and repeated after inoculation till the "taps"
became bloody and unfit for differential counting. Considerable
variation in the percentage of lymphocytes occurred in the
inoculated mice, and also some variation in the same mouse at
different times. The lowest differential count recorded showed
25 per cent lymphocytes, the highest 80 per cent of lympho-

PERITONEAL IMPLANTATION OF SARCOMA

cytes, and the average in uninoculated mice 56 per cent lympho-
cytes. After inoculation, when the tumors were well developed
and readily palpable through the skin, there was found an
average of 54 per cent of lymphocytes. Twenty-two mice were
used in this experiment. About half the animals showed a rise
and the other half a fall in the percentage of lymphocytes fol-
lowing tumor inoculation. Table 1 illustrates the results

TABLE 1

PERCENTAGE OP LYMPHOCYTES

MOUSE

TUMOR

Before inoculation

After inoculation

Large

None

Large

Of 22 mice, 11 show an increase in lymphocytes, and 11 a decrease.

obtained before and after inoculation. Following the inocula-
tion, numerous polymorphonuclear leucocytes appeared in the
peritoneum but not in sufficiently large numbers to influence
visibly the percentages of the other cellular elements.

Thus, as indicated in the table, of 22 mice inoculated with
tumor all with one exception presented in a short time (ten to
twenty days) rapidly growing tumor nodules. In making the

ELLIS KELLERT

punctures with capillary glass tubing blood was frequently
obtained and when abundant the slides were discarded. The'
counts recorded above were made on relatively blood-free
preparations. Since many mice showed an increase and others
a decrease in the numbers of lymphocytes, and since the tumors
were large and growing rapidly, it is apparent that no marked
adverse influence can be attributed to the presence of the lympho-
cytes. The most important point brought out by the present
investigation is the fact that small fragments of tumor tissue
placed in a fluid rich in lymphoid cells showed no inhibition
of growth.

CONCLUSIONS

1 . The normal peritoneum of the white mouse is rich in cellular
elements — 115,000 to the cubic millimeter — of which about 55
per cent are lymphocytes.

2. Mouse tumor implanted into the peritoneum of these mice
grows rapidly.

3. No marked change in the cellular content of the peritoneum
of such mice results from the inoculation of mouse tumor.

4. No direct antagonistic action is found to exist between the
lymphocyte and tumor implants in mice.

REFERENCES

(1) Da Fano: Ztschr. f. Immunitatsforsch., Orig., 1910, v, 1.

(2) Burgess: Jour. Med. Res., 1909, N.S., xvi, 575.

(3) v. Dungern and Coca: Ztschr. f. Immunitatsforsch., Orig., 1909, ii, 391.

(4) Apolant: Ztschr. f. Immunitatsforsch., Orig., 1913, xvii, 219.

(5) Tyzzer: Jour. Med. Res., 1915, N. S., xxvii, 201.

(6) Braunstein: Quoted by Oser and Pribram, Amer. Jour. Med. Sci., 1913,

cxlvi, 310.

(7) Rohdenburg and Johnston: Quoted by Oser and Pribram, Amer. Jour.

Med. Sci., 1913, cxlvi, 310.

(8) Baeslack: Ztschr. f. Immunitatsforsch., Orig., 1914, xx, 421.

(9) Murphy: Jour. Exper. Med., 1913, xvii, 482.

(10) Mur phy and Morton: Jour. Exper. Med., 1915, xxii, 204, 800.

(11) Wade: Jour. Path, and Bacteriol., 1908, xii, 384.

(12) Rous and Murphy: Jour. Exper. Med., 1912, xv, 270.

(13) Serafini: Tumori, 1917, v, 229.

(14) Jones and Rous: Jour. Exper. Med., 1914, xx, 404.

PERITONEAL IMPLANTATION OF SARCOMA

(15) Bullock: Lancet, London, 1915, i, 701.

(16) Stevenson: Jour. Cancer Res., 1917, ii, 245.

(17) Sittenfield : Jour. Cancer Res., 1917, ii, 151.

(18) Stevenson: Jour. Cancer Res., 1917, ii, 449.

(19) Bullock and Rohdenburg: Jour. Cancer Res., 1917, ii, 465.

(20) Bullock and Rohdenburg: Jour. Cancer Res., 1918, iii, 31.

(21) Prime: Jour. Cancer Res., 1920, v, 105.

(22) Evans: Amer. Jour. Physiol., 1915, xxxvii, 243.

PLATE 1

Figs. 1 and 2. Examples of rapidly growing peritoneal tumors following tumor
inoculation into mice with numerous lymphocytes in the peritoneum.

PERITONEAL IMPLANTATION OF SARCOMA

ELLIS KELLERT

Fig. 1

4«.i

PLATE 2

Figs. 3 and 4. Peritoneal smears from normal mice. Obj. 16 mm. Oc. 5.
Note the large numbers of lymphocytes. The larger pale-staining cells are macro-
phages. A few mast cells are present.

PERITONEAL IMPLANTATION OF SARCOMA

ELLIS KELLER

PLATE 2

Fig. 3

Fig. 4

PLATE 3

Fig. 5. Peritoneal smear showing large numbers of macrophages. Tumor
growth rapid. Obj. 4 mm. Oc. 5.

Fig. 6. Peritoneal smear from mouse which showed 75 per cent of lympho-
cytes in presence of well-developed peritoneal tumor. Obj. 4 mm. Oc. 5.

PERITONEAL IMPLANTATION OF SARCOMA
ellil'kellert

PLATE 3

•A.: •

- * • t '.%

Fig.

Fig. 6

PLATE 4

Fig. 7. Peritoneal smear showing 75 per cent lymphocytes. Tumor large and
rapidly growing. Obj. 4 mm. Oc. 5.

Fig. S. Peritoneal smear showing 70 per cent lymphocytes. Tumor large.
Obj. 4 mm. Oc. 5.

PERITONEAL IMPLANTATION OF SARCOMA

ELLIS KELLERT

u • •••• • *

• • •• • •

. ' '• ": •

A « •

Fig. 7

PRIMARY SPONTANEOUS SQUAMOUS CELL
CARCINOMAS IN MICE

STUDIES ON THE INCIDENCE AND INHERITABILITY OF
SPONTANEOUS TUMORS IN MICE

Fifteenth Communication

MAUD SLYE, HARRIET F. HOLMES, and H. GIDEON WELLS

From the Otho S. A. Sprague Memorial Institute and the Department of Pathology
of the University of Chicago

Received for publication April 4, 1921

Throughout most of the animal kingdom primary carcinoma
of the skin is met with but rarely, in contrast with its great
frequency in man, and also with the relative frequency of other
cutaneous and subcutaneous tumors in animals. The existing
literature on spontaneous tumors in animals indicates that no
species except dogs, horses, and fowls, show skin carcinoma
except as a rare condition, and this is of particular significance
in view of the fact that skin carcinomas are more certain to be
recognized than those in any other location and hence then-
relative infrequency is even more marked than any statistical
evidence would indicate. Presumably the unprotected con-
dition of the human skin accounts for its susceptibility to cancer,
as the hairy scalp rarely shows primary carcinoma, and the
pigmented skin of the negro also seems insusceptible.

The classical compilation of Sticker (1), which contains records
of many tumors not fully differentiated according to modern
criteria, gives the following relative frequency in this material
of skin tumors in different species :

58 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

SPECIES

Horse

Cattle

Dog

Cat

Swine

Total carcinomas

332

766

Skin

166

Eye

Lip.*

Anus

Penis

Vulva

The extensive compilation by Teutschlaender (2) of the distri-
bution of tumors in animals shows that cases of squamous cell
carcinoma have been described in but few species, according to
his records, namely: cat, dog, mouse, rat, elephant, beef, sheep,
pig, horse, mule, cockatoo, toucan, chicken, carp, and gold fish.
However, in another table he mentions also a case in a stag,

Apparently birds develop cutaneous carcinoma somewhat
more often than most other animals, but even here there are
but few recorded. In their summary of the literature on tumors
in birds, Joest and Ernesti (3) found but 4 of 37 tumors described
in birds to be skin carcinomas, those being two found on the legs
of chickens by Wernicke; one on the wing of a "papapeis" —
reported by Guerrini, and one on the wing of a toucan reported
by Herbert Fox (4), this last case exhibiting also a metastasis
in the lung. In none of their own 50 cases of tumors in birds
was a skin carcinoma described. This compilation, as well as
all others that we have observed, omits the striking case of
squamous carcinoma of the skin of the foot of a hen described
by Boynton (5), in which transplantations were made without
success. More recently a similar tumor arising on the thumb
claw of a lark has been reported by Urra (6). Two cases of
squamous cell carcinoma of the mouth in fowls have been
described by Pick and Koch. The experience of Teutschlaender
(7) is unique, for he reports that of 54 carcinomas in fowls
observed in the cancer laboratory at Heidelberg, 28 were skin
epitheliomas, of which 18 were on the foot, attributed to the
irritation of parasitic infection of the epidermis {Cnemidocop-
tesmilben) .

SQUAMOUS CELL CARCINOMAS IN MICE

Apparently fish are subject to skin tumors of various structure,
including carcinoma. Thus, Fiebiger (8) reports cases of epi-
thelioma occurring on the lip of each of two "Schlei" {Tinea
vulgaris Cuvier) coming from the same pond, a markedly malig-
nant skin epithelioma in a carp, and mentions papillomatous
growths in several specimens of Kletterfish (Anabus scandens).
Plehn has also described numerous skin epitheliomas in cyprinoid
fishes.

Bashford has reported a squamous cell carcinoma in a small
fish, Gasterosteur spinachia, two cases of carcinoma of the skin
in frogs, and a carcinoma of the skin glands in a Triton.

Among mammals the dog alone seems to be commonly sub-
jected to skin carcinoma, if we except the carcinoma of the penis
and prepuce occurring not infrequently in horses and oxen. It
would seem to be not uncommon in cats according to the experi-
ence of Roffo (9), who found 7 tumors in 307 cats examined, of
which 4 were skin carcinomas on the head and face, with lymph
gland metastases, but none of 11 cases of tumors in cats col-
lected by Bashford (10) was of this type, and but 2 of 48 tumors
in dogs, although there were several squamous cell growths in
the mouth and pharynx. Leo Loeb (11) reported in 1903 that
from one to three carcinomas of the inner canthus of the eye
were observed every week in the Chicago stock yards, but in
one year no other external carcinomas were observed except
two of the vulva.

Rodents seem to have skin carcinoma but rarely. The large
series of wild rats autopsied in plague work and examined for
tumors (McCoy, Woolley and Wherry, Beatti) have revealed
no such growths except for an epithelioma of the vulva reported
by Woolley and Wherry (12). It will be recalled that the early
transplantation experiments of Hanau (13) were with a squamous
cell carcinoma from the vulva of a white rat, and he states that
there had been two previous cases of similar growths in rats in
their laboratory. Roffo (14) has also observed a squamous cell
carcinoma which arose in the vulva of an old white rat, which was
transplanted through six generations without change in structure.
In view of the great frequency of parasitic skin infection in rats,

60 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

"rat scabies/' commonly with extreme papillomatous prolifer-
ation of the epithelium (15), it is strange that epithelioma has
not been more frequently observed. Of 123 cases of spontaneous
rat tumors reported in the literature and collected by Bullock
and Rohdenburg (16), there were but three epitheliomas, one of
the tongue and two of the vulva. They reported on 48 rat
tumors observed in the Crocker Laboratory, one of which was
an epithelioma of the head, which was not described further.
Morris (17) has reported a basal cell carcinoma which arose in
the skin of a male rat, five months old, and which died out after
two generations of transplants. He was unable to find other
recorded instances of basal cell carcinomas in animals.

Several instances of squamous cell carcinoma have been
described in mice, but they form an insignificant proportion of
the many spontaneous tumors that have been observed in this
species. Nevertheless they were looked upon as of great signifi-
cance in the early days of experimental cancer research, since
they established the fact that mice had other forms of tumor
than the predominating mammary gland carcinoma. Haaland
(18) reported a case of squamous cell carcinoma arising in the
mouth of a mouse in BorrePs laboratory, involving the lower
jaw, showing a typical squamous cell structure. Six mice were
inoculated from this, but as in six months they showed no
growths they were put back into the cage with the others. Ten
months after the inoculations "deux cas identiques de cette
meme tumeur apparurent en meme temps dans cette cage,"
but it was not known whether these were inoculated mice or not
and Haaland does not make clear whether these later tumors
were in the mouth, although this is apparently what he means.
Later he observed a third case of mouth cancer in a mouse
recently inoculated with a Jensen tumor. No growths developed
in mice inoculated from this mouth tumor.

Murray (19) in 1908 described the following cases of super-
ficial squamous cell carcinoma in mice: (1) A prickle cell growth
with apparently little keratinization, arising in the skin of the
neck, with metastasis into an adjacent lymph gland. (2) A
carcinoma presenting both alveolar and squamous cell areas,

SQUAMOUS CELL CARCINOMAS IN MICE

with keratinizing metastases in the lung. Presumably this tumor
arose in the mammary gland. (3) Prickle cell growth without
keratinization, interpreted as primary in the nipple; this mouse
also had had a hemorrhagic adenocarcinoma of the mammary
gland removed by operation. In discussing the mammary
gland tumors he also describes the occurrence of areas of keratin-
ization within tubular carcinomas, and regards them as indi-
cations of the close association of the mammary apparatus with
the skin from which it develops, and points out that the ampulla
which receives the terminal portions of the mammary ducts is
also lined by stratified squamous epithelium. "Therefore, should
the cells of the new growth have taken their origin near the
nipple, variations in either direction are only to be expected."

In his later report from the London laboratories, Haaland
(20) states that of 353 spontaneous tumors observed (not
including those in Murray's report), 22 were squamous-celled
carcinomas with marked keratinization, of which 14 arose in
the mammary region and 8 outside it. Of these 8, 3 were mouth
tumors similar to those observed in Bon-el's laboratory, appar-
ently arising in the mucous membrane of the inside of the cheek,
there being in one metastases in the lymph nodes. One seemed
to have arisen on the surface of the head, and it penetrated the
skull; grafts from this were successful. (The illustrations of
these tumors show the gross relations of such tumors so well
that we are spared the necessity of reproducing further illus-
trations.) Three tumors were described as superficial verrucous
growths, on the vulva, anus, and left flank. One was a typical
squamous cell carcinoma of the skin of the chest, developing as
a superficial ulcerating growth, "undoubtedly from the skin
itself, or from the mammilla." A case of mixed squamous and
sebaceous carcinoma of the vulva is also mentioned.

Several of the keratinizing mammary tumors differ from
typical squamous cell carcinomas, and their cutaneous origin is
doubtful. Four are described as resembling "molluscoid" tumors,
characterized by long radiating cylinders of cells with a central
keratinized zone, found in the mammary region and not giving
successful transplants. Four tumors are described as "typical

62 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

squamous-celled alveolar carcinomata in the mammary region. "
Six others are called " adenocancroids, " these being unques-
tionably mammar}^ gland adenocarcinomas in which areas of
keratinization occur. Attention is called to the occurrence of
squamous-celled cysts in the normal mamma, apparently arising
from mammary ducts in which the epithelium has become
metaplastic, and which perhaps explains the occurrence of
keratinizing carcinoma in the mammary gland. As squamous-
celled growths have also been found rarely in the human breast,
as well as in cats and dogs — Teutschlaender (21) — the disease
in mice is not without analogy, although apparently relatively
frequent among the mammary gland tumors of this species.
Presumably in some stocks of mice keratinization is even more
common in mammary gland carcinoma than the above reports
indicate, for Woglom (30) found this change in 228 of 1000
spontaneous mammary gland carcinomas examined in the Crocker
Laboratory.

Pick and Poll (22) reported as a "sweat gland carcinoma" a
tubular growth arising near the scapula of a mouse, but the
description published does not permit the exclusion of a mammary
gland origin.

Erdheim (23) described a pedunculated tumor, which gave no
evidence of malignancy, but exhibited a histological structure
resembling squamous cell carcinoma. It arose from the vulva
of a mouse, and after the outer part of the growth had been
removed the base healed spontaneously. He was unable to
classify this tumor, beyond indicating that it was a stratified
epithelial neoplasm.

Tsutsui (24) has produced carcinoma on the skin of the back
of mice by painting with tar, after the procedure used by Yama-
giwa with the ears of rabbits, observing 16 carcinomas and 1
sarcoma in 17 mice surviving over 100 days. In two cases lung
metastases were observed.

We find no other reports of squamous cell carcinoma arising
in the skin or mouth of mice, which indicates their relative
infrequency, since such tumors are most obvious when they do
occur. Thus, Tyzzer (25) in his report of 83 spontaneous

SQUAMOUS CELL CARCINOMAS IN MICE

tumors in mice includes no tumors of this type. In the 41
primary mouse tumors described by Jobling (26) there were no
skin epitheliomas, although there was one of the molluscoid
keratinizing tumors such as Murray, Haaland, and Tyzzer have
described.

Squamous cell carcinomas outside the mammary gland and
skin of mice are rare. Tyzzer reported that among his lung
tumors "in several cases" the growth was of an epidermoid
character, and Haaland described one such tumor. In our
previous paper on primary carcinoma of the stomach in mice
(27), we collected four cases of squamous cell gastric carci-
noma reported in the literature and added three more in our
own material, as well as one case of squamous carcinoma in the
external surface of a chronically prolapsed rectum. Fiebiger
(31) has also reported the occurrence of a few instances of
squamous cell carcinoma produced experimentally in the stomach
of mice by feeding cockroaches injected with Spiroptera. We
have found no reports of squamous cell carcinomas arising in the
esophagus or urinary tract, or in the cervix uteri, or in any other
structure where they might be found in mice, except such as
have been mentioned above.

OBSERVATIONS ON SKIN TUMORS AND SQUAMOUS CELL
CARCINOMAS IN THE SLYE STOCK OF MICE

In 28,000 consecutive autopsies performed on mice of this
stock, which had been permitted to live as long as possible
without any experimental manipulations whatever, we have
observed the following instances of primary neoplasms of squa-
mous or stratified epithelial structure.

Primary squamous cell carcinoma of the skin and mouth 70

Primary basal cell carcinoma of the skin 15

Primary keratinizing carcinoma of the mammary gland 56

Primary squamous cell carcinoma of the stomach 4

Primary keratinizing carcinoma of the lung 1

Primary squamous cell carcinoma of the rectum 2

Primary squamous cell carcinoma of the vagina 1

Primary stratified carcinoma of the meibomian gland 2

Primary sebaceous gland adenocarcinoma 1

64 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

As we have not yet made a complete analysis of all the tumors
observed in these 28,000 autopsies the proportion of stratified
cell tumors to total tumors cannot be stated. Roughly there
are probably about 4000 primary spontaneous tumors of all sorts,
so that these squamous cell tumors constitute not far from 4 per
cent of the total.

CARCINOMA OF THE SKIN

In our material are 85 malignant epithelial growths arising in
the skin, mouth, and lips, excluding those squamous cell growths
that seem to be derived from mammary gland tissue and strat-
ified cell growths from other glands. Of these, 70 are of squamous
cell type and 15 of basal cell character. Since many of the
tumors which arise about the lips and mouth are, when observed,
so extensive that their exact point of origin cannot be deter-
mined, we have grouped them together with the skin tumors.
Roughly classifying the site of the 70 squamous cell growths, 13
were in the skin of the trunk, one on a front limb, 2 on the vulva,
15 about the lower jaw, 18 on some other portion of the face,
and 23 about the ears or neck; that is, all but 16 of the 70 were
on the head and neck. Presumably this is to be explained by the
much greater amount of traumatism suffered by the skin in this
part of the body, which is especially marked in cage mice which
are always rubbing their muzzles against the rough wire meshes
of the cages. The tumors of the jaw and mouth in several
instances seemed to have resulted from the irritation produced
by broken or protruding teeth. Nearly all of the carcinomas
arising on the trunk were definitely located at the site of a healed
wound, and in not a few of those of the head the same origin
was observed ; it is probable that less obvious wounds are respon-
sible for many if not all of the others. A chronic dermatitis
often preceded the carcinomas of the skin of the face, presumably
incited by traumatism. The not uncommon fungus infections
of the skin seem to be too rapid in their course to lead to carci-
noma formation.

SQUAMOUS CELL CARCINOMAS IN MICE

Clinically these growths are similar to corresponding forms of
carcinoma in man (fig. 1). The age at which they occur is, on
an average, later than with any other mouse tumors, as they
rarely appear before late middle life, and usually only in old
age. They arise under crusted ulcers, or develop in areas of
hyperplasia from chronic irritation, spread slowly as crusting,
ulcerated growths, destroying at times the eyes or other features,
usually remain superficial, and most often cause death by chronic
infection of the ulcerated surface, which may also lead to acute

Fig. 1. Head and Face of Mouse with Extensive Basal Cell Carcinoma
of the Skin, of at Least Five Months Duration. No. 7874

abscess formation. Occasionally death results from starvation
when the cancer involves the mouth or jaw, from hemorrhage,
invasion of the skull, or other coincident diseases.

The microscopic diagnosis of the skin neoplasms is often
difficult, especially with the basal cell growths, since we find all
stages of epithelial overgrowth from simple hyperplasia to
metastasizing carcinomas. In this series no growth is included
as a carcinoma unless it exhibited both the gross and clinical
features of malignancy together with distinct microscopic
evidence of infiltrative character. The extensive destruction of
these growths by infective processes often obscures the micro-
scopic findings. In several instances a growth that presented

Fig. 2. Typical Squamous Cell Carcinoma of the Jaw, with Formation of
Abundant Epithelial Pearls. No. 23331. X 60

Via. 3. Squamous ( Jell ( Iarcinoma of the Jaw Infiltrating Adjacent Tissues
Extensively. Type with Little Hornifn ltion. No. 26238. X 60

SQUAMOUS CELL CARCINOMAS IN MICE

all the clinical features of carcinoma has been found at autopsy
so extensively necrotized and suppurating that only examina-
tion of numerous sections from different parts of the growth
has established the diagnosis of carcinoma. It is quite possible
that we have lost a few genuine carcinomas through such destruc-
tive processes, to say nothing of cases in which the cannibal-
istic mate has selected the neoplasm for the first course.

Fig. 4. Squamous Cell Carcinoma of Skin of Chest Arising in the >
of an Old Wound, the Scar Tissue Being Abundant in this Field,
and the Relation of the Cancer Growth to the Over-
lying Skin Being Shown. No. 8212. X 45

The microscopic features of the squamous cell carcinoma
the skin and mouth of mice differ not at all from those that
occur in man. Usually horniflcation is marked (fig. 2), but, as
in man, growths from a similar origin may show little or no
tendency to form keratin (fig. 3). Usually the amount of
connective tissue formation is not large, except in some cases
where the cancer has developed in scar tissue from old wounds
(fig. 4). Little help in diagnosis is afforded by mitotic figures,

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

since these are scant}' in even the most typically malignant'
growths, and commonly cannot be found at all. Presumably
they would be much more numerous if the tumors were removed
from the living animal.

The infiltrative character of growth is not usually so extensive
as we are accustomed to see it in man, for the mice generally
succumb to infection from the ulcerated growth while it is

Fig. 5. Squamous Cell Carcinoma of the Head and Face Invading the
Spinal Column in the Cervical Region. Compressing the
Spinal Cord and Infiltrating the Meninges
This tumor also invaded the skull. No. 23766. X 60

still small. Infiltration is usually seen besl in the tumors that
arise about the jaws, for here the bone is often invaded. Bone
infiltration was observed in 7 of the L5 tumors that arose about
the jaw. We have had one striking ease (237(H)) in which a
carcinoma of the skin of the head, beginning at the base of the
right ear, infiltrated the skull and cervical vertebral canal,
infiltrating the meninges about the cerebellum and compressing
the spinal cord in the cervical region (fig. 5). This case resembles

SQUAMOUS CELL CARCINOMAS IN MICE

the one described and pictured by Haaland, in which the skull
was invaded. In two instances we have seen infiltration of the
salivary glands, and in three cases an adjacent lymph node
was involved apparently by direct extension.

The absence of lymphatic metastasis is a striking feature
of these tumors when compared with corresponding growths in
man, which is true of all forms of carcinoma in mice. In only
two cases did we find a secondary growth by metastasis into a
lymph node from a skin carcinoma (7950, 15232), and it was a
very common observation that lymph nodes immediately
adjacent to or in contact with these squamous cell growths,
were not involved. Possibly serial sections of all our mice would
have revealed other instances of metastasis, but the value of
the information did not seem commensurate with the labor
involved. In only one case did we find visceral metastasis
(12627). This mouse had a squamous cell carcinoma arising
just dorsal to the rectum, and a typical secondary nodule about
1 mm. in diameter in the lung; it also had a tubular carcinoma
of the mammary gland which had produced no metastasis. The
infrequency of metastasis in this series of spontaneous squamous
cell carcinoma is significant when compared with Yamagiwa's
experimental tar tumors in rabbits, since the infrequency of
metastasis in his material has been thought b}r some to speak
against their being true malignant tumors. As a matter of fact
the metastasis incidence obtained by him, and by Tsutsui with
experimental skin tumors in mice, is distinctly higher than that
observed in these spontaneous growths.

The sex incidence is strongly in favor of the female (49 to 21)
in this series. As most of the growths were on the head and
neck we cannot account for this on the basis of erroneous inclusion
of squamous carcinomas of the mammary gland with our skin
tumors. It differs from our experience with other growths of non-
reproductive tissues which have shown approximate equality as
to sex incidence. Even more than with other tumors we have
found that age is an important factor, these squamous cell carci-
nomas of the skin being predominatingly in old mice, and as
many male mice die early from wounds received in fighting

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

this may account for the relative preponderance of females in
this series.

Beyond the relatively slight extent of infiltration and metas-
tasis these carcinomas present no noteworthy differences from
human skin carcinomas. Often the amount of keratin scales piled
up on the surface is strikingly great, and in one case the carci-
noma arose at the base of definite cutaneous horns (25785).
There is often a noteworthy amount of calcification of the
necrotic scales. Cyst formation is frequently observed, and
benign cutaneous cysts have been observed in several mice.

The basal cell carcinomas all arose on the face, ears, and head
(fig. 1), 9 of the 15 being in females. Because these growths
are of relatively low malignancy, as in man, they present partic-
ular difficulty in diagnosis, and we have set aside as " precan-
cerous" numerous instances of basal cell hyperplasia of marked
degree which did not present unqualified proof of malignancy.
Quite frequently enormous hyperplasia of the cells about the
hair follicles produces tumors of considerable size, and in these
may be found areas highly suggestive of malignancy, in that
altogether atypical plugs and masses of basal cells are formed,
as shown in figure 6. Such growths, which very probably
would have shown unqualified malignancy had the mouse lived
longer, have not been included among the basal cell carcinomas.
In general, basal cell carcinoma in the mouse corresponds entirely
to the corresponding growth in man (fig. 7). To quote Mac-
Callum (28), "In spite of the complexity of the downward
growing strands all reach to about the same level. Further,
it is seen that they are very sharply outlined against the stroma
and show little inclination to strew their cells into the irregular
crevices of that tissue." We have found the various modifica-
tion- of basal cell growths commonly described, such as the
formation of tubules suggestive of undeveloped hair follicles,
hornifying surface areas with typical basal cell growths beneath,
and the so-called adenoid epithelioma which seems intermediate
between basal and squamous cell growths. Among our keratin-
izing carcinomas of the mammary gland was one of cells
that suggested a basal cell character. No metastases were

Fig. 6. Basal Cell Hyperplasia of an Irritated Muzzle, with Areas of
More Atypical Growth
There have been numerous such growths, which are probably properly desig-
nated as precancerous, but may be actually malignant. Xo. 16563. X 60.

g. 7. Basal Cell Carcinoma Which was Highly Malignant, as It
Entirely Replaced the Face, Destroying Both Eyes, and Forming
Tumor Masses in the Orbits. No. 9932. X 90

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

observed among these cases, although in one, extensive infiltra-
tion destroyed both eyes and produced large tumor masses in
the eye sockets. Several of these growths arose at the site of
distinct scars from wounds.

Despite the fact that these skin tumor mice were all well
advanced in age the co-existence of other tumors is lower than
that seen in many other forms of mouse tumor. Perhaps one
factor in this is that most of the mice with skin cancer had lived
beyond the age at which other tumors occur most frequently.
It also seems that the heredity of these mice is responsible to
some extent, but a complete analysis of this factor has not yet
been made. Of the 85 cases, in which 58 were females and 27
males, but 6 showed a carcinoma of the mammary gland, 7
exhibited lung adenoma (including one that also had a mammary
gland carcinoma), one had an adenoma of the ovary and one
seemed to have pseudoleukemia. Despite the number of old
wounds in these mice no cases of sarcoma were observed, presum-
ably because of either old age or ancestry, but one case of carci-
noma arising in the mouth and infiltrating the jawbone, showed
such a flattening of the deeper cells that for some time we were
in doubt whether or not it was an instance of mixed sarcoma
and carcinoma. This case (8560) is described and pictured in
our article on sarcoma in mice (29).

SQUAMOUS CELL CARCINOMAS OF THE MAMMARY GLAXI)

The keratinizing carcinomas of the mammary gland form an
interesting group, and properly lie outside the scope of this
paper, except for the fact that with not a few of them it is
extremely difficult to tell whether we are dealing with a primary
mammary gland carcinoma or a skin carcinoma arising over the
mammary gland. Three types of these tumors can be dis-
tinguished.

1. Carcinomas of the mammary gland which are essentially
cylindrical cell carcinomas, forming tubules and alveoli, but
some areas of which undergo a transformation into stratified
epithelium with the formation of keratin, often in large amounts
(fig. Sj. These are the "adenocancroids" of Murray and

Fig. 8. Tubular Alveolar Carcinoma (" Adenocancroid") of the Mammary
Gland with Areas of Keratinization
To the left is seen a band of densely keratinized tissue with no evidence of
its glandular origin. No. 13336. X 60.

Fig. 9. Superficial Adenocancroid, Apparently Arising in the Ducts of

the Mammary Gland
Shows very little evidence of origin in glandular tissue. This mouse also
had three other tubular carcinomas of the mammary gland without keratiniza-
tion No. 15300 X 60.

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

Haaland. It is probably of some significance that the keratin-
ization is usually most marked in the portions nearest the
cutaneous surface, as if it began in the ducts. The amount of
keratinization varies, sometimes appearing in only a few small
spots in the tumor tissue, but often extending until little of the
original columnar cell type of tissue remains.

Fig. 10. Keratinizing Cysts in the Mammary Gland
Such structures are not infrequently found, and may be the precursors of some
of the malignant keratinizing carcinomas of the mammary gland. No. 8271.
X 40.

2. Squamous-celled keratinizing carcinomas without any
evidence whatever of cylindrical cell structure, but arising
Bubcutaneously in the mammary regions (fig. 9). Many of these
present no histological evidence that they are derived from the
cells of the mammary gland, but they are observed to arise
within the gland substance and sometimes are still entirely
subcutaneous when the mouse dies. The occasional presence

SQUAMOUS CELL CARCINOMAS IN MICE

within the mammary gland of what seem to be simple benign
cysts of stratified epithelium with masses of desquamated horni-
fied material (fig. 10) indicates the probable origin of such tumors.
When they ulcerate on the surface it may be impossible to
differentiate them from primary carcinoma of the skin or nipple,
for histologically they differ little if at all from the usual squamous

Fig. 11. Molluscoid Carcinoma of the Mammary Gland

This growth, which arose in an inguinal mammary gland, invaded the abdomi-
nal wall and protruded into the pelvis, but produced no metastases. A non-
keratinizing carcinoma of the mammary gland was also present which produced
pulmonary metastases. No. 15622. X 60.

cell carcinoma of the skin. When observed early their origin
in the gland beneath the skin is usually the chief ground for
recognizing them as mammary gland tumors; nevertheless, the
fact that after differentiating, on a histological basis solely, the
skin carcinomas from the mammary gland carcinomas we found
that nearly all those we had selected as of cutaneous origin had

76 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

arisen at the site of old skin wounds, indicates that there are
usually recognizable differences.

3. Molluscoid carcinomas of the mammary gland, to adopt
Borrel's designation of these striking growths with long, radia-
ting cylinders of stratified epithelium with a large central core of
keratinized material (fig. 11). These seem to grow out from the
tubules near the nipple, as if they arose from the stratified
portion of those ducts. They produce ulceration of the surface
about the nipple, which greatly enhances their resemblance to
the skin carcinomas.

We have little to add to the statements of Borrel, Murray,
Tyzzer, and Haaland concerning these squamous cell growths
arising within or about the mammary gland, and the excellent
illustrations accompanying Haaland's article make any further
illustrations by us unnecessary. A few points may be worth
recording. Forty-three of the 56 squamous-celled mammary
gland tumors may be best designated as of the adenocancroid
type, in that they show a greater or less proportion of cylindrical
cell structure. Several of these were multiple, as is often the
case with primary carcinoma of the mammary gland. In three
there were two adenocancroids, one mouse had three, and one
had five, four arising distinctly in mammary gland tissue while
the fifth arose anterior to the urethra and seemed to have origi-
nated in the skin, and this showed no glandular elements. Four-
teen mice had an adenocancroid and also one or more other
independent mammary gland carcinomas that showed no keratin-
ization. Of these, one (9365) exhibited also a typical squamous
cell carcinoma metastasis in the lung; one mouse with three
primary adenocancroids (5738) showed a lung metastasis while
three mice that had each two mammary gland tumors, one of
each type, had pulmonary metastasis of the cylindrical cell type.
Thai is, from 43 cases of adenocancroid of the mammary gland,
but one metastatic nodule of adenocancroid has been seen in
the lungs, and none in the lymph nodes or elsewhere. Not a
few of these mice had also co-existing tumors other than those of
the mammary gland, namely, one with a squamous cell carci-
noma of the muzzle (9962), one with a malignant ovarian tumor

SQUAMOUS CELL CARCINOMAS IN MICE

(6801), two with uterine fibroids (14370 and 14811); one with a
lung adenoma (12098) ; one with an adenoma of the liver and an
adenoma of the lung (9544) ; one with an adenoma in the lung
and a strange growth in a subcutaneous lymph node that might
be interpreted as an endothelioma (15622) and one had an intra-
pelvic sarcoma with metastasis into the liver and also into an
ovary which contained in addition a small adenoma (10006).

Two of the adenocancroids were examples of the not uncommon
mixture of sarcoma and carcinoma in the mammary gland of the
mouse (15412 and 6182).

Only four of our tumors were typical of the molluscoid cancers,
although more or less of this character was shown by several of
the growths diagnosed as adenocancroids. One of these four
mice had also an adenoma of the ovary (496) and another had a
benign lung adenoma (5581). None of these four keratinizing
tumors produced metastases.

Nine squamous cell carcinomas that exhibited no glandular
structure to identify them as mammary gland tumors, were
believed to have this origin because they arose beneath the skin,
and in most instances did not ulcerate through at any time.
These growths, however, seem to be of a more malignant character
than the adenocancroids; three of them produced squamous
cell metastases. No. 10936, which is remarkable in being the
only male with a subcutaneous stratified epithelial growth that
seemed to arise in the mammary gland, exhibited squamous
cell metastatic growths in the lung, in the mediastinum, and in the
chest wall. Another (7950), with three subcutaneous, non-
glandular squamous cell carcinomas, showed a metastatic nodule
in one lymph node and direct infiltration of a second. No.
13671 with a subcutaneous squamous cell carcinoma and also a
cylindrical cell mammary gland carcinoma, had three large
metastatic nodules in the lungs all of squamous cell structure.
As to multiplicity of tumors in this group there were three
independent subcutaneous squamous cell tumors in each of two
mice (2239 and 7905), two had a simple cylindrical cell carcinoma
of the mammary gland in addition to the squamous cell growth
(7526 and 13671), while one had adenomas of the liver and ovary

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

(14047) and one mouse had leukemia (8236). In this group
we have included one squamous cell tumor (24858) that differs
from the rest in having the non-keratinized tumor elements
suggest strongly the appearance of a basal skin carcinoma,
although the growth seemed to arise in the mammary gland.

SQUAMOUS CELL CARCINOMA OF STOMACH AND RECTUM

In a previous paper on primary carcinoma of the stomach
in mice (27) we collected records of four reported cases of squa-
mous cell carcinoma arising in the cardiac portion of the stomach
to which we added three observed in this laboratory. Since
that publication we have observed one more case (24367). This
was an old female mouse that for some time before death had
been isolated because suffering from tape worm infection. The
stomach wall in the cardiac portion was much thickened and
small white nodules were scattered over the outer surface. No
metastases could be found. In the liver was an encysted cestode,
and there was a prolapsed rectum showing much thickening of
the wall. Microscopically the nodules in the stomach are of
squamous cell carcinoma, apparently still quite early but of
typical structure. One small nodule of similar structure is found
in the omentum. The rectum shows a marked overgrowth of
squamous epithelial plugs on the outer, ulcerated surface of the
bowel, and with many of these an infiltrative character is sug-
gested, but apparently this lesion cannot be diagnosed as
carcinoma.

Another new case of carcinoma of the stomach (25911) is more
difficult to classify. A female Peromyscus mouse, nearly four
and a half years old, which had been isolated all her life and
never bred, refused food for several days before death. At
autopsy there was found a marked thickening of the entire
stomach wall, particularly at the cardiac end where it seemed
nearly solid from wall to wall. A thick white nodular outgrowth
6 by 4 by 4 mm. of similar tissue bound together the stomach,
liver, and esophagus, while another nodule 8 by 4 by 4 mm.
bound together the pancreas and intestinal mesentery posterior

SQUAMOUS CELL CARCINOMAS IN MICE

to the stomach. No other metastases were found. Microscop-
ically this growth is composed of groups of strands of infiltrating
epithelial cells, which are not hornifled, neither do they form
tubules. They grow profusely throughout all coats of the
stomach, sometimes forming sheaths about the tubules of the
cardiac portion. The nodules outside the stomach are secondary
growths of identically the same structure in lymph nodes. It
is not possible to be sure whether this carcinoma is derived from
the columnar or the stratified epithelial portions of the stomach.
While the gross appearances suggested an origin in the cardia,
which is the site of all but one of the recorded cases of gastric
carcinoma in mice, yet all of these were frankly hornifying
squamous cell carcinomas.

The full details of the three other cases of squamous cell
carcinoma occurring in this series are published with illustrations
in our previous communication (27) and need not be repeated
here. To recapitulate them briefly they were as follows:

No. 5802, male two years old. Squamous cell carcinoma of
pyloric portion of the stomach, measuring 15 by 15 by 12 mm.
A secondary growth 11 by 12 mm. in the mesentery. No other
metastases. Some infiltration of the pancreas.

No. 7851. Male, twenty-five months old. Cardiac portion
of the stomach is ulcerated and thickened, forming a mass 16
by 12 by 8 mm. No metastasis. Structure, typical squamous
carcinoma, infiltrating all coats of the stomach, and invading
the adhesions between the stomach and liver.

No. 16440. Female, aged twenty-seven months. At the
junction of the cardiac and pyloric portions a thickened mass
from 3 to 15 mm. wide surrounds the stomach. Nodules were
found in the omentum and mesentery, metastases in lymph
nodes. Structure, typical squamous cell carcinoma with some
hornification.

In the same paper was published with illustrations the report
of a case of squamous cell carcinoma arising in the metaplastic
epithelium covering a prolapsed rectum (8345) . This was a male
mouse, which had had a prolapsed rectum for six months before
its death. The growth was not extensive but seemed to be

80 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

typically squamous cell carcinoma arising in metaplastic epi-
thelium. Since then we have observed a second similar case.

No. 20052. Female, had a prolapsed rectum for four months
before death, which was from hypertrophy of the heart and
pulmonary edema. The surface of the prolapse became much
ulcerated and after a time showed what seemed to be a prolif-
eration of epithelium. The regional lymph nodes were some-
what enlarged. Microscopically the external surface of the
prolapsed bowel shows, in addition to considerable ulceration,
areas of definite squamous cell carcinoma, with strands of cancer
cells infiltrating through a thick layer of granulation tissue
down to but not into the muscularis. Apparently this has
arisen from what was part of the cutaneous surface of the anus,
and not from metaplastic mucosa as in the previous case. As
a possible factor in the inciting irritation, a small piece of wood
was found imbedded in the bowel wall, surrounded by granula-
tion tissue and cancer cells.

CARCINOMA OF THE VULVA AND VAGINA

No cases of squamous cell carcinoma of the uterus have as
yet been described in mice, so far as we can learn. Although
we have found a few fibromyomas and sarcomas of the uterus,
we have met with but one epithelial neoplasm, which was an
adenocarcinoma.

Erdheim (23) and Haaland (20) each have described a recurrent
verrucous growth of the vulva, and the latter a mixed squamous
and sebaceous cell carcinoma of the vulva. Two cases of
squamous carcinoma of the vulva have been observed in this
series and are included among the 70 squamous cell skin carci-
nomas. (1) No. 7950. This occurred as a condylomatous
growth arising distinctly in the vulva of an old mouse, infiltra-
ting the subcutaneous tissues, and was an unquestionably
malignant squamous cell growth microscopically. There was in
addition metastatic growth in two lymph nodes in the groin.
(2) No. 18928. A growth of a warty character and ulcerating
slightly developed on the vulva some time before the death

SQUAMOUS CELL CARCINOMAS IN MICE

of the mouse, spreading about the rectum before death. At
autopsy there was found in addition an enlarged lymph node
attached to the right ureter. Microscopically the external
growth is a typical squamous cell carcinoma with much horni-
fication, and not very marked tendency to infiltration. The
nodule attached to the ureter was a lymph node containing a
cyst lined with squamous epithelium, or, to describe it better,
an epithelial cyst covered with a thin layer of lymphoid tissue.
The epithelial wall of the cyst is thin, shows no evidence of infil-
trative or other malignant character, and hence does not at all
resemble a metastatic growth, but it is difficult to explain the
presence of such a structure in this location on any other basis.

In addition to these two carcinomas we have observed an
excellent case of carcinoma of the vagina.

Carcinoma of vagina (no. 22582). This mouse was found,
nearly a month before its death, with a 3 mm. pink nodule
protruding from the vagina. At the time of its death from
pneumonia it presented an ulcerating mass 15 by 12 by 10 mm.,
about one-half of which protruded from the vulva. It seemed
to arise from the vaginal wall, the uterus and bladder not being
involved, but it ulcerated into the rectum and was much infected
and ulcerated. No enlarged glands or other evidences of metas-
tasis could be found. Microscopically the growth is composed
mostly of loose masses of keratinized scales exfoliated from the
underlying growth, which infiltrates the vaginal wall as a typical
squamous cell carcinoma.

KERATINIZING TUMORS OF THE LUNG

As mentioned in our review of the literature, Haaland has
described one case of primary keratinizing growth in the lung,
and Tyzzer says that he has had " several cases. " The influence
of heredity on the incidence of tumors of special types is suggested
by the fact that of several hundred cases of primary lung tumors
in the Slye stock, but one has been definitely found to show
keratinization.

No. 13314. A male, age twenty months, which showed no
other autopsy findings of interest, had the lower lobe of the left

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

lung nearly replaced by a yellowish mass, which was distinguished
from the ordinary papillary adenoma of the lung chiefly by its
color. This made it resemble an abscess, but it was much too
hard to be an abscess, and there was no pleural exudate or
adhesion. Microscopically the tumor consists chiefly of a mass
of hornified scales heaped up in waving, concentric layers.
Only at the very periphery are living cells found. Here is a
narrow growing border of stratified epithelial cells, differing in
no essentials from that seen in epitheliomas of the skin. There
is no marked tendency to infiltration, the growth apparently
progressing by expansion, but the presence of occasional mitotic
figures is noted. About the growth there is much round and
spindle cell proliferation and numerous foreign body giant cells
are found. Our specimen differs from the one illustrated by
Tyzzer in having a smaller proportion of living cells. We are
not certain whether this growth represents a true neoplasm or a
progressive metaplasia due to some persisting chronic inflam-
matory condition, but the former seems more probable.

Two other mice have shown somewhat related pulmonary
conditions.

No. 10561. This mouse, which had also a carcinoma of the
skin, had a benign adenomatous growth in the lung which showed
some tendency to stratification, but without keratinization.

No. 25136. A male mouse had in the right upper lobe of the
lung a mass 14 by 12 by 10 mm. resembling in appearance a
malignant tumor of the lung. Microscopically this tissue
resembles much more closely an unresolved organizing and
necrotizing pneumonia, in which are two irregular, independent
areas composed of masses of keratinized scales, with a slender
border of flattened epithelial cells. It is quite impossible to
decide whether this is a true tumor, or whether it is the cause or
the result of the pneumonic condition, although the epithelial
growth has the appearance of being much older than the pneu-
monic process.

SQUAMOUS CELL CARCINOMAS IN MICE

TUMORS OF THE MEIBOMIAN GLAND

Two mice have presented growths arising in the eyelids,
which, according to their structure, seem to be adenomas arising
in the Meibomian glands.

No. 18099. A small, slowly growing mass developed beneath
the left eye of a female mouse; from the e}re exuded a small
amount of thick white exudate. The growth had reached a
diameter of 10 mm. when the mouse died from an acute lung
infection, and showed no evidence of infiltration or ulceration.
Microscopically the growth is composed of papillary structures
covered with many layers of epithelial cells. It differs from the
normal Meibomian gland in the exaggeration and lawless arrange-
ment of the structures, and the greatly increased number of
epithelial cells covering the stroma. The diagnosis of benign
adenoma of the Meibomian gland seems justified, especially in
view of the size of the growth and the findings in the next case.

No. 27929. An old female mouse, which died of chronic
nephritis, had a soft mass, 6 mm. in diameter, beneath the left
eye. The gross appearance suggested an epithelioma. Micro-
scopically this tumor is quite the same as the one described
immediately above, except for the important fact that it infil-
trates down to the bone of the orbit, thus indicating that it is a
malignant infiltrating adenocarcinoma.

SEBACEOUS GLAND ADENOCARCINOMA

Such tumors have been described in mice by Murray, Tyzzer,
and Haaland, the last two having transplanted them success-
fully. We have found one case of sebaceous adenocarcinoma of
the preputial gland, which closely resembles the growth described
and illustrated by Haaland (20).

No. 19895. Beginning at the base of the penis is a mass 30
by 25 by 25 mm. extending well into the inguinal region. The
penis was completely imbedded in the tumor, the older portions
of which were softened, but about the periphery were hard
nodules of newer growth. The testicles and epididymis were

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

not involved. Death resulted from chronic nephritis. There
were no metastases. Microscopically this tumor (fig. 12) repro-
duces closely the normal structure of the preputial gland, but
not infrequent infiltration of the stroma by strands of epithelial
cells corroborates the gross evidences of malignancy. It corre-
sponds perfectly to the illustration given by Haaland, who also

Fig. 12. Sebaceous Adexo-carcixoma of Preputial Glaxo

This was a large tumor with distinctly malignant characteristics, infiltrating
the adjacent tissues, despite its close resemblance to the normal structure of
the gland from which it arose. Xo. 10695. X 60.

comments on the close resemblance to the normal gland in spite
of the definitely malignant character.

SUMMARY

Among 2S,()()() mice dying natural deaths at all ages, and
carefully autopsied, have been observed growths of strati-
lied and squamous epithelium that correspond by the usual

SQUAMOUS CELL CARCINOMAS IN MICE

standards to true neoplasms, excluding a considerable number
of epithelial growths which lack positive conclusive evidence of
neoplastic character, although possibly some of these are also
early carcinomas. Seventy-one are examples of squamous cell
carcinoma of the skin or mouth. They differ from the human
skin carcinoma chiefly in a low incidence of metastasis. Fifteen
others are of basal cell character, arose always about the head,
and produced no metastases. In both these groups the incidence
is higher in the females than in the males. Trauma and chronic
irritation seem to play an important part in the production of
skin carcinoma in mice, most of our cases occurring about the
head and face, often recognizably at the site of wounds, and
nearly all the skin carcinomas of the trunk arose in old scars.
Skin cancers occur at a greater average age than other tumors
in mice.

Fifty-six examples of squamous cell keratinizing growths
arising in the mammary gland were observed, predominatingly
adenocarcinomas with localized areas of keratinization. These
also seldom produce squamous cell metastases.

Other tumors in this group were: Four squamous cell carci-
nomas of the stomach, two arising in the prolapsed rectum, two
in the vulva, one keratinizing tumor of the lung, one sebaceous
adenocarcinoma of the preputial gland, and, as hitherto unde-
scribed mouse tumors, one squamous carcinoma of the vagina
and two adenomas of the Meibomian glands, one of these being
infiltrative and apparently malignant.

The literature of the comparative pathology of squamous
cell carcinoma in animals is reviewed, and it is worthy of comment
that as yet no cases of squamous cell carcinoma of the uterus,
bladder, or esophagus seem to have been described in mice.

REFERENCES

(1) Sticker: Arch. f. klin. Chir., 1902, lxv, 616, 1023.

(2) Teutschlaender: Ztschr. f. Krebsforsch., 1920, xvii, 285.

(3) Joest and Ernesti: Ztschr. f. Krebsforsch., 1915, xv, 1.

(4) Fox: Jour. Path, and Bacteriol., 1912, xvii, 217.

(5) Boynton: Festschrift for Victor C. Vaughan, 1903, p. 535.

(6) Urra: Rev. de med. y cirug. pract., Madrid, 1918, cxviii, 321.

MAUD SLYE, H. F. HOLMES AND H. G. WELLS

(7) Teutschlaender: Ztschr. f. Krebsforsch., 1920, xvii, 350.

(8) Fiebiger: Ztschr. f. Krebsforsch., 1909, vii, 165.

(9) Roffo: Rev. d. Inst. Bacterid., Buenos Aires, 1918, 1, 333.

(10) Bashford: Third Sci. Rep., Imperial Cancer Research Fund, London

1908, p. 41.

(11) Loeb: Arch. f. klin. Chir., 1903, lxx, 845.

(12) Woolley and Wherry: Jour. Med. Res., 1911, xxv, 205.

(13) Hanatt: Fortschr. d. Med., 1889, vii, 321.

(14) Roffo: Rev. d. Inst. Bacterid., Buenos Aires, 1918, i, 155.

(15) Lipschutz: Wien. klin. Wchnschr., 1920, xxxiii, 426.

(16) Bullock and Rohdenburg: Jour. Cancer Res., 1917, ii, 39.

(17) Morris: Jour. Cancer Res., 1920, v, 147.

(18) Haaland: Ann. de l'lnst. Pasteur, 1905, xix, 165.

(19) Murray: Third Sci. Rep., Imperial Cancer Research Fund, London, 1908

p. 69.

(20) Haaland: Fourth Sci. Rep., Imperial Cancer Research Fund, London

1911, p. 1.

(21) Teutschlaender: Ztschr. f. Krebsforsch., 1920, xvii, 300.

(22) Pick and Poll: Berl. klin. Wchsnchr., 1903, xl, 518.

(23) Erdheim: Ztschr. f. Krebsforsch., 1906, iv, 33.

(24) Tsutsui: Gann, 1918, xii, 17.

(25) Tyzzer: Jour. Med. Res., 1909, xxi, 479.

(26) Jobling: Rockefeller Inst. Monographs, 1910, i, 81.

(27) Slye, Holmes and Wells: Jour. Cancer Res., 1917, ii, 401.

(28) MacCallum: Text-book of Pathology, Philadelphia, 1920, p. 1032.

(29) Slye, Holmes, and Wells: Jour. Cancer Res., 1917, ii, 1. Fig. 3.

(30) Woglom: Jour. Cancer Res., 1917, ii, 471.

(31) Fiebiger: Jour. Cancer Res., 1919, iv, 367.

PROCEEDINGS OF THE AMERICAN ASSOCIATION FOR
CANCER RESEARCH

THIRTEENTH ANNUAL MEETING
Held in New York, April 1, 1920

1. Report of the Council

The meeting of the Council was held at Dr. Wood's apartment in
New York City, on the evening of Wednesday, March 31, 1920.

The following members were present: Dr. H. Gideon Wells, presi-
dent; Dr. Robert B. Greenough, vice-president; Dr. Francis C. Wood,
and Dr. Willy Meyer, and, by invitation Dr. Frederick Prime. Absent,
Dr. E. R. LeCount, Dr. James B. Murphy, and Dr. William H. Woglom.

The report of the treasurer for the year showed a balance on hand
of $532.27.

The status of the Journal of Cancer Research was discussed,
and a statement from the publishers, showing a total deficit of $110.80
on the four volumes so far published, was read. After some discussion
of the proposal, made by the publishers, that the price of the Journal
be increased to $6.00 for subscribers not members of the Association,
it was voted that no change be made until after the completion of the
publication of Volume V.

In order to increase the membership of the Society it was decided
that the president should send out a letter to certain men whose
names were presented as possible candidates for active and associate
membership. •

The name of the following applicant came before the Council for
election to the Association:

Dr. Michael Levine, Montefiore Home, New York City.

Dr. Wood moved and Dr. Wells seconded the motion that he be
elected to membership in the Association.

The resignations of the following members were accepted:

Dr. J. George Adami, Dr. Martha Tracy.

The deaths of Dr. M. J. Herzog and Dr. James Douglas were noted.

Dr. Wood moved that Dr. James Ewing be elected councillor suc-
ceeding the retiring councillor. Dr. Greenough seconded this motion.

The following officers were elected by the Council to serve for the
ensuing year. Dr. Robert B. Greenough, president; Dr. James B.
Murphy, vice-president; Dr. William H. Woglom, secretary and
treasurer (re-elected).

THE JOURNAL OP CANCER RESEARCH, VOL. VI, NO. 1

88 PROCEEDINGS

The present Council, therefore, with the years of retirement, is as
follows :

Dr. E. R. LeCount, 1921 Dr. William H. Woglom, 1924

Dr. F. C. Wood, 1922 Dr. Robert B. Greenough, 1925

Dr. James B. Murphy, 1923 Dr. Willy Meyer, 1926

Dr. James Ewing, 1927

The Council continued in office the present Editorial Board, which
is composed as follows:

Dr. William H. Woglom Dr. Leo Loeb

Columbia University Washington University

Dr. Frederick Prime Dr. Ernest E. Tyzzer

Columbia University Harvard University

Dr. Joseph C. Bloodgood Dr. H. Gideon Wells

Johns Hopkins University University of Chicago

Dr. James Ewing
Cornell University

2. A Few Late Results After the Radical Operation for
Cancer of the Breast

Dr. Willy Meyer (New York) :

SUMMARY

The data given are from the author's private cases which he has
followed for the last twenty-six years.

Two radical operations for cancer of the breast have been before the
profession since the fall of 1894. Their principal point of difference is
the direction in which the surgeon proceeds. One method starts from
the chest and works toward the axilla, leaving the clavicular portion
of the pectoralis major behind; it requires entering the space between
the pectoralis major and minor muscles, the latter , usually being
divided and then sutured. This method necessarily involves loss of
blood. The other method, practised by the author since September
12, 1892, starts from the axilla and works toward the sternum. The
tendons of the pectoralis major and minor are divided in the early
stage of the operation, necessitating complete excision of both muscles.
Blood- and lymph-vessels are primarily divided within the axilla.
The lymph-nodes and axillary fat are lifted out in connection with
the tumor, before the cancerous breast itself is handled. The entire
mass is removed without entering what he calls the "infected area."
Hemorrhage is reduced to a minimum.

The final results of the operation from the sternum toward the
shoulder, as reported, have been good. Still, small cancerous nodes
have been found repeatedly between the pectoralis major and minor
muscle, and where cancerous lymphatic nodes have developed, there
must be present suspicious lymphatic vessels.

PROCEEDINGS

Previous to 1894, excision of the breast for carcinoma was done in
two stages, but at the same sitting, first, the removal of the breast with
axillary contents; then, the excision of the pectoralis major muscle.
This arrangement forced the surgeon to enter widely "the infected
area" and caused an unnecessarily great loss of blood. Meyer did not
see a single lasting cure after this method of advance. The radical
operation changed the results with one stroke, as is shown by a com-
parison of the author's statistics before and after September, 1894.
The first two patients subjected to the modern radical operation were
completely cured, and case 4 of this series had for many years enjoyed
freedom from cancerous recurrence when she died of old age.

The following patients in the series are alive and well today, from
twelve to twenty-five and a half years after operation:

Case 1. Operation in September, 1894; patient thirty-eight years of age
(now sixty -four). This was the first case operated upon by the method out-
lined above. The patient is alive and well today, twenty-five and a half years
after operation.

Case 2. Operation in 1895; patient forty-eight years of age (now seventy -
three). This woman is alive and well today, twenty-five years after operation.

Case 3. Operation in July, 1902; patient thirty-three years of age (now fifty-
one). The patient is alive and well today, eighteen years after operation.

Case 4- Operation in December, 1903; patient thirty-six years of age (now
fifty-three). The patient is perfectly healthy and free from recurrence today,
seventeen years after operation.

All these patients have full use of the arm and are able to assume the posture
of the "Statue of Liberty."

Case 5. Operation in July, 1908; patient thirty-five years of age (now forty-
seven). She is perfectly healthy today, twelve years after operation.

Case 6. This patient was operated on in September, 1917, and the case is
added merely to show the present line of incision with Handley's addition down
to a point midway between the umbilicus and the xyphoid process, for the ex-
cision of the fascia covering the upper portion of the recti muscles, in conjunc-
tion with the other mass. This addition is recommended because it makes the
operation more radical and usually makes possible closure of the wound, without
grafting.

Five other patients remained free from recurrence for 4, 6, 8 (2
cases), and 16 years, respectively, and then died of other diseases.

Another patient, a pronounced diabetic at the time of the operation,
was well for six years after it, when she succumbed to diabetes, without
having developed any signs of a recurrence of the cancer.

Still another patient, operated on for cancer of the right breast in
March, 1899, returned in December, 1900, with a carcinoma of the
left breast, which was then also extirpated. She was well and free
from recurrence when last heard from, in the spring of 1907, six and a
quarter years after the second operation.

A few days ago, the author saw a patient now almost eighty years
old and in perfect health, who had been operated upon by him for a
scirrhus of the breast at the age of seventy-three (seven years ago).

These results, he believes, prove the efficiency of the method; they
prove that the radical operation for cancer of the breast can cure patients

PROCEEDINGS

thus afflicted. That not all the patients are saved is due (1) to the
stage of the disease in which they reach the surgeon, and (2) to the
virulence of the agent that produces the carcinoma.

The importance of the follow-up system was discussed in connection
with the collection of statistics on the late results of operation.

PageVs disease {epithelioma of the nipple)

Paget's disease is the most malignant of known forms of cancers of
the breast. If ever early and radical operation is imperative, it is in
these cases, as will be seen from the following three observations which
were made during the last two years.

Case 1. Female, thirty years old, mother of five children, had been in the
hands of a quack and had been treated by caustics. When seen in January, 1918,
the disease in the breast and axilla had far advanced. After the radical opera-
tion the other breast soon became affected, and one year later, also was excised.
Then, not long after, the disease became disseminated, and the patient died
from cancer en cuirasse.

Case 2. Female, thirty-eight years old, had been in the hands of one of our
best x-ray specialists in the city. One and one-half years after cure of the dis-
ease of the nipple by radium treatment, there was a local recurrence and a very
extensive cancer of the breast, with infected nodes in the axilla and along the
subclavian vein. Radical operation was done in December, 1918, followed by
renewed x-ray and radium treatment. She now has developed intrathoracic
metastases.

Case 3. Male, forty-seven years old, had been in the hands of an experienced
surgeon who had extirpated the breast only, without the axillary nodes, evidently
because none could be found at that time. One and three-quarter years later,
the patient presented a far advanced carcinoma. The radical operation then
performed could not save him; and he died from general metastases eight months
afterward.

In operating upon mammary carcinoma, Meyer makes it a point to
circumcise the skin widely at the base of the breast. He prepares two
ample flaps and enfolds them extensively, then divides the fasciae at
the base of the two flaps, and extirpates them together with the mass.
Involvement of the supraclavicular nodes does not, in his opinion,
present a contraindication to operation; on the contraiy, he considers
it the surgeon's duty to operate when these nodes are infected. Hal-
sted, as well as the late Rodman, have observed patients who remained
well for a number of years after the extirpation of these nodes.

In none of the cases presented were the supraclavicular nodes found
infected at the time of operation, and hence they were not removed.
Meyer has not operated on a single case in which there were no infil-
trated axillary nodes.

Meyer believes that statistics regarding the results of the radical
operation for cancer of the breast are worthless. They do not prove
anything. What does determine the fate of the patients is the so-
called virulence of the disease. One and the same surgeon may do an
equally radical operation in two seemingly early or apparently equally
far advanced cases, in one of which the patient may remain well and

PROCEEDINGS

free from recurrence for, say, twenty-five years, while in the other a
regional recurrence and metastases may develop within a few months.
All that can be said is that cancer, being a local disease in the begin-
ning, may be cured by a radical operation, if this is done at an early
stage.

DISCUSSION

Dr. Robert B. Greenough (Boston) ; I feel we must acknowledge that
surgery of the breast owes much to Dr. Meyer for the work he has
been doing, and I am entirely in accord with his reasons for operating
in the way which he recommends, namely, beginning the operation in
the axilla, and removing the axillary contents, the whole of the breast,
both of the pectoral muscles, and the fascia of the chest wall, all in one
piece. I feel this hardly needs argument, because it is an accepted
principle in operating for the radical cure of cancer in any situation
where the anatomical conditions make it possible. I personally have
had more experience with the transverse axillary incision than with
the incision used by Dr. Meyer, but I think the choice of the incision is
a very small matter. It is far more important what is done beneath
the skin, and I am entirely in accord with Dr. Meyer in that respect.

In regard to the presentation of statistics in cases operated on for
the cure of cancer, some fourteen years ago I had the opportunity of
reporting, at the meeting of the American Surgical Association, the
results of operations for carcinoma of the breast at the Massachusetts
General Hospital, and at the same meeting other papers on the end
results of cancer at many different institutions were also presented.
Dr. Halsted opened the meeting, and the very first words of his paper
were of great significance. He said: "It is especially true of breast
cancer that the surgeon interested in furnishing the best statistics
may in perfectly honorable ways provide them."1 I believe that if
we are going to attempt a comparison of different methods of operating
we must have some definite standard which we accept. I know no
better standard than to take all the cases that come into a general
hospital under the diagnosis of the disease in question and to record
them during the period under observation, with a statement as to what
percentage of that number was considered suitable for the attempt at
radical cure, and what results were obtained. I am sorry that the idea
has not taken hold better in the general reporting of cases. I was
looking at a report of cases of cancer of the lip recently published, and
in that report all the cases in which an attempt to trace the patient
after operation by letter had been unsuccessful, were wiped out of the
record. It has been our experience in Boston that the patients from
whom one cannot hear are usually those that have actually left this
world for another, and I fully believe that patients untraced should be
counted as failures. I believe also, as Dr. Meyer stated so emphati-

1 Halsted, W. S. : The Results of Radical Operations for the Cure of Carci-
noma of the Breast, Ann. Surg., 1907, xlvi, 1.

PROCEEDINGS

cally, that one of the great functions of the social service departments
of these hospitals should be the tracing of end results. At the Hunt-
ington Hospital we have an excellent social service worker who has
established a standard in regard to these things, and wTe do not con-
sider a record closed when the patient has been discharged until we
have definite information of the end result. We keep them as live
cases until we know that they are dead.

3. Primary Spontaneous Tumors of the Ovary in Mice — Studies
in the Incidence and Inheritability of Spontaneous
Tumors in Mice. Fourteenth Report

Miss Maud Slye, Miss Harriet F. Holmes, and Dr. H. Gideon Wells
(Chicago) :

SUMMARY

Among 22,000 mice of the Slye stock dying natural deaths at all
ages were 44 with spontaneous primary ovarian tumors, not including
simple ovarian cysts. Of these, 38 had simple benign solid papillary
adenomas, only occasionally with slight cyst formation; 1 showed a
typical papillary cystoma, and 1 a typical solid teratoma containing a
groat diversity of tissue elements. A second case of this type has
since been found in the first 25,000. Of the 38 cases of solid papillary
adenomas, 19, or 50 per cent were bilateral, so that there were 57
tumors of this class. There were 4 unquestionable primary malignant
tumors of the ovary, all showing the " mesothelioma" type of growth
characteristic of malignant tumors derived from the sex glands; one of
these produced perirenal metastases. One other tumor of the same
type was primary in either the ovary or the adrenal. Two round-cell
sarcomas were found, arising either from the ovary or from some other
organ, while 2 other sarcomas had produced secondary growths in the
ovary. Of the 44 mice with primary ovarian tumors, 26 had tumors
in other parts of the body.

In the literature were found reports of eight other cases of benign
tumors arising in the ovaries of mice, all exhibiting the same character-
istics as the tumors described in this paper.

discussion

Dr. F. C. Wood (New York): Statistics of this sort are of vital impor-
tance in all experimental work in cancer, since t hey show that if animals
are kept to a sufficient old age there is a very large incidence of tumors
of all varieties. I think that we who experiment with animal tumors
are still supposed by the clinical fraternity in medicine to be working
with something which is entirely different from and not in the least
comparable to human tumors. But as observations on animals are
being extended over larger series of mice, and to other species, as white
rats, dogs, guinea-pigs, etc., we are finding instead that tumors abso-

PROCEEDINGS

lutely comparable in morphology and biological qualities occur in
many of our domesticated animals. This is interesting as furnishing
an argument against the current opinion that tumors are a disease of
civilization, and primarily of the educated classes, and not a disease
which is widespread and generalized through all groups of society.
The statistics of the Metropolitan Life Insurance Company, which
show a greater incidence of cancer among the laboring classes, point
in the same direction, that is, to the fact that there is no immune
class. Obviously, in the manual laboring classes the conditions of
irritation and of infection (for instance, syphilis, which, while not
causing cancer, facilitates its occurrence) result in a larger proportion
of tumors in that group, and as we study mice we find that larger
numbers of tumors spontaneously appear in these animals. I do not
doubt that the same would be found to be true of wild mice if we could
keep larger numbers to old age.

Dr. James W. Jobling (New York) : I should like to ask Dr. Wells if
these ovarian tumors were observed more frequently among the " tumor
strains" of mice described by Miss Slye.

Dr. Wells: In reply to Dr. Jobling's question: Miss Slye had hoped
to be able to discuss that feature, as, of course, the study of heredity
is entirely her part of the work; but on account of illness she was not
able to get the material together. We have had strains for twenty-five
or thirty generations producing enormous numbers of mice with no
tumors. It is perfectly safe to say that mice of certain strains are
more likely to have tumors of the ovary than are those of other strains.
It is a familiar fact that in the strain of mouse tumor that is used in
most laboratories in America, that derived from the Abbie Lathrop
stock in Granby, Massachusetts, cancer of the mammary gland is the
usual type. Very few cases of tumors of any other tissue have been
described in this stock; but Haaland's reports indicate that in the
mice in the Imperial Cancer Research Fund in London, tumors of
other organs are quite common. I may recall Miss Slye's observa-
tions, previously reported, that with tumors of specific organs a very
definite relationship to heredity is shown. She reported at one time
twenty-eight cases of primary tumor of the testicle, all, with one
exception, arising in mice of one definite strain. The exception was
in a mouse which had been bitten on the testicle, and which belonged
to a strain in which sarcoma was very common. The most striking
series which Miss Slye has had was the liver tumor. You may recall
that the literature of mouse tumors had previously shown but one
case of primary tumor of the liver; since then one or two others have
been described. Now Miss Slye has bred, from a mouse with primary
tumor of the liver, a strain in which there have developed one hundred
cases of primary growth in this organ. While there are no figures on
the ovarian tumors, we know that most of them arise in certain strains,
though whether strictly within the strains, as in the case of the testicle
and liver tumors, we are not yet prepared to say. But in such experi-

PROCEEDINGS

ments as these, certain difficulties must be faced. Thus Tyzzer and
Haaland have described squamous-cell neoplasms of the lung, a type
of which we have seen only one or two examples. Hence minute
differences must exist, making one strain liable to one variety of tumor
in any given organ, and another liable to a type slightly different.

Dr. C. C. Little (Cold Spring Harbor, N. Y.) : How would you explain,
on the basis of a single Mendelian factor, the occurrence of these clearly
demarcated strains which show specialized types of tumors in particular
organs or groups of organs?

Dr. Wells: I am not prepared to discuss that aspect of the work at
all; that I leave entirely to Miss Slye. I do not consider myself com-
petent to discuss the matter of genetics in these problems.

Dr. Little: I would like to point out that the result outlined by Dr.
Wells is very interesting in support of the theory that the hereditary
nature or susceptibility to the occurrence of spontaneous tumors
depends on more than one hereditary factor. This is evidenced by the
clear ability of a particular family to localize its susceptibility to tumor
in a particular organ, and is further evidenced by the ability of other
families to combine tumors of different organs. Such a result is par-
ticularly characteristic of the action of many factors in heredity, rather
than of only one. I mention the fact because of the possibility that
too much emphasis may be placed on a single factor as the cause of all
cancer. The specificity of tissues is so great that it would be surpris-
ing if such evidence as that given by Dr. Wells were not found when
so careful a study was made.

Dr. William C. Stone (New York): I am very much interested in
this presentation because of my own study of ovarian tumors in the
human subject. In the first place, as to diagnosis: Dr. Wells spoke of a
certain number of these tumors as possibly endotheliomata. Similarly
in the human cases, in going over the literature of the so-called Kru-
kenberg tumors, one finds numerous instances in which the picture has
been interpreted by many as an endothelioma, by others as a sarcoma,
and by still others as a carcinoma. Then, as regards the observation
of the occurrence of the tumors in both ovaries, and the difficulty of
explaining this in human patients, also, in a large number of instances
the ovarian tumors are of secondary nature, and the primary site is
elsewhere; but there are numerous cases in which no primary site can
be found, and the incidence of the tumor in both ovaries is unexplained.
There is one difference, however, if I understand Dr. Wells correctly,
between human and mouse tumors. In the majority of cases the so-
called Krukenberg tumors in women seem to be secondary in the
ovaries. If one considers carefully the descriptions of those reported
as primary, one is unable to exclude the possibility that in the major-
ity there was a primary site elsewhere. We found it most frequently

PKOCEEDINGS

in the stomach, gall-bladder, appendix, or some other part of the
intestinal tract.

Dr. Wells: Of course, in these cases the histology is not that of the
typical Krukenberg tumor. The tumors are bilateral, but they do
not resemble the Krukenberg tumor. They do show characteristics
which I am quite sure would be cause for a diagnosis of endothelioma
or of sarcoma; in fact, before I had studied these tumors carefully I
labelled a good many of them endotheliomata of the ovary. It may be
interesting to know that every once in a while we see in bovines bi-
lateral tumors of the ovary; some of those I have seen from the stock-
yards in Chicago are histologically similar to the Krukenberg tumor
although nothing is found to indicate a primary tumor elsewhere in
the cattle. In mice there have been reported only five or six cases of
primary abdominal tumors which could give rise to Krukenberg
tumors; so that here we have tumors which are bilateral, and which are
distinctly not secondary tumors. Why they should be bilateral in 50
per cent of the cases, and why bilateral tumors are so much more likely
to occur in the ovaries I am unable to explain.

4. The Role of Neoplasia in Parasitic Diseases of Plants.
Dr. Isaac Levin and Dr. Michael Levine (New York) :

Previous investigation by the authors on the crown gall have demon-
strated that while this condition frequently acts in a manner analogous
to animal cancer, the cellular proliferation is primarily a reaction to
the invasion of Bacterium tumefaciens. The previous experiments
were conducted on annuals, biennials, or deciduous trees, in which the
period of growth of the host as well as of the crown gall is normally
interrupted. In these experiments, some of the galls are benign to
the host and behave in a manner more analogous to a scar, a cheloid,
than to cancer. In a comparatively small percentage of cases the galls
act as true malignant tumors. The parts of the inoculated stem
become necrotic above and even below the point of inoculation.

The present investigation was conducted on the rubber tree (Ficus
elastica), which is an evergreen perennial plant and grows indoors, so
that the crown galls may be watched for long periods of time without
interference by secondary contaminations. These experiments have
shown that twelve months and more after inoculation, nearly every
crown gall produces a necrosis of the inoculated branch, though at
first the gall may attain a large size without apparent injury to the
host-branch. Then frequently, without any additional increase in
the size of the gall, the branch becomes necrotic both below and above
the gall. This necrosis increases in a centrifugal direction from the
gall; thus the tip of the branch may still be alive while a part of the
branch below the gall is necrotic.

These phenomena cannot be due to impairment of nutrition since
the crown gall frequently attains its largest size without a concomitant

PROCEEDINGS

necrosis, and the latter is always subsequent; nor can the necrosis be
due to the action of some toxic substances produced by the crown-gall
cells. It is difficult to conceive of such a rapid change in the metabolic
chemical functions of a cell. The most plausible explanation is a
change in the reactivity of the host tissue.

For a time the Bacterium tumefaciens produces a progressive reaction
in the injured and surrounding tissues of the host, which manifests
itself in proliferation of cells and formation of a crown. Sooner or
later this is followed by regressive reaction which manifests itself in
progressive necrosis. Since plants lack the lymphoid tissues and can-
not react to parasitic invasion by inflammation, the above described
methods of reaction are the only possible ones.

Thus, the neoplasia in the crown-gall disease is primarily a protective
reaction of the host tissue to the invasion by Bacterium tumefaciens.
The morphological studies by the writers of two other parasitic dis-
eases, clubroot of cabbage (Plasmodiophora brassicce), and potato-wart
disease, or potato cancer (caused by Chrysophlyctis endobiotica) show
that identical reactive cell proliferations and formation of new growths
may be caused by other parasites besides Bacterium tumefaciens.

It is possible that in most parasitic diseases of plants both neoplasia
and necrosis take place, though the former may be so insignificant and
transitory that it evades detection.

DISCUSSION

Dr. Wells: It is pleasant to hear that potatoes and cabbages contain
items of interest besides calories and vitamines. This subject is, of
course, of great importance, because the extremely interesting work of
Dr. Smith has attracted so much attention, and there has been a lively
discussion as to whether these processes in plants are really to be con-
sidered as true neoplasms. I wish to express my personal apprecia-
tion of this contribution because it must be determined whether these
things in plants are true tumors or not, since their interpretation will
have much bearing on our evaluation of the work in cancer research.

5. A Phase of Tumor Biology
Dr. Frederick Prime (New York):

SUMMARY

In the past few years much confusion has arisen on account of the
various results reached by different cancer investigators doing prac-
tically the same line of research. A great deal of this confusion is
probably due to the use of tumors whose biological characteristics are
really unknown to the investigator. Applications are made to this
laboratory every year by workers on the cancer problem who desire a
carcinoma or sarcoma which they wish to propagate for a short time in

PROCEEDINGS

order to carry out some investigation. They then report their results
on a few dozen animals, whereas if the characteristic action of the
tumor over numerous generations had been known to them, or if they
had used a larger number of animals their results would have been
very different. For instance, the Jensen rat sarcoma is reported in
one paper to take in almost 100 per cent of animals inoculated. In
running back over our books for the last six years we find that in 5400
animals surviving at the end of three weeks we have only one series
in which there was 100 per cent of takes, and only a few in which the
takes were as high as 90 per cent. If our experiment had been done
in March, 1914, for instance, we should have had 100 per cent takes
in our controls, whereas if it had been done in July, 1916, we should
have had only 20 per cent positives. In another instance important
conclusions were drawn from two series of tumor 63, one with a high
percentage of takes, the other with no takes, but our books show that
the same fluctuation has occurred in routine transplantation. The
results vary from month to month and from year to year, but if the
averages for the months are made the variations are less marked. Of
all our tumors, the mouse sarcoma Crocker Fund No. 180 gives the
highest percentage of takes and is the most consistent, the number of
takes never having fallen below 85 per cent. The Ehrlich mouse
sarcoma is a close second to this and in only one month did this fall
below 80 per cent of takes.

Certain members of our staff, I myself among them, have had the
general impression that our tumors grew less well during the summer
months, but upon analysis this proved to be quite incorrect; the death
rate among the animals may have been higher, but the tumor growth
rate did not show any such change. For six years the Jensen rat
sarcoma had its highest number of takes in August, and its lowest in
July. The Flexner rat carcinoma had the highest number of takes in
April, and the lowest in November. It behooves us, therefore, to
know and study our tumors carefully before starting any investigations
on them, and to recognize the fallacy of using small numbers of animals
in arriving at a conclusion.

DISCUSSION

Dr. S. R. Benedict (New York): I think it would be interesting if
Dr. Prime could give some more definite statements as to how much
these fluctuations would influence conclusions, that is, whether the
results would be diametrically opposed through an increased number of
animals, and whether he would recommend that an experiment be
based on 5000, or on 500, or on 50 animals. A general criticism of this
type calling attention to the number of animals might be misinter-
preted unless it is made more specific. In the matter of transplanting
a given tumor, for instance, we have quite exact data, and if we trans-
plant from an experimental animal, using the same tumor, we should
get a similar growth in both series, except for the experimental factor.

PROCEEDINGS

I should like to ask how large a series would have to be, and whether
it must be repeated for several months before definite conclusions can
be drawn.

' Dr. Prime: I think the criticism is exceedingly pertinent. We feel
that a good deal of confusion arises with men working along the same
line who report such varying results from exactly the same tumor,
because they do not quite know the biology of the tumors. One may
report results with a certain series or a certain tumor and another give
data of an opposite nature, but that does not necessarily mean that
one or the other is incorrect. If a little more were known about that
tumor, it would probably be found that neither was making very great
errors. I think one advantage of a laboratory of this kind is that we
have tumors in such large series, and that our records are always avail-
able, so that anyone can refer to a tumor with which he is working.
But so many of the people who come here for tumors do not care what
they get or what happens to it. They want a tumor they can grow
in one or two generations. I do not think conclusions should be drawn
from experiments with 25 or 50 animals. I think each series should
contain one or two hundred animals, and that one series alone should
not be taken as a basis for conclusions. As you see here our results
vary from month to month; and if experiments were made on several
series and the results averaged, I think they would be very much more
valuable than are deductions drawn from one or two small groups, as
is often done. It is simply that we want to avoid the error of making
too sweeping conclusions from too few animals and from a lack of
knowledge of the growth rates of these tumors. I think that eventually
if experiments are made in this country with tumors whose growth and
behavior is known, and with animals whose behavior is known, results
from the various laboratories would be very much more consistent.

Dr. Wood: The question has other phases, also, such as the statistical
study of tumor regression after taking and the study of the tumor
growth rate. In some instances the number of takes is most impor-
tant, for instance, when attempting to establish immunity against a
tumor it is vital that we know that the tumor is not one which estab-
lishes immunity against itself. Some workers have reported experi-
ments in which they claimed to have established immunity by means
of treatment of the animals, but they overlooked the fact that the
tumor used was one which immunized against itself. In experiments
in which attempts are made to influence the growth of a tumor by
therapeutic means, it is necessary to use a tumor in which the number
of spontaneous disappearances is negligible. Our no. 180 is such a
tumor, and because of this fact and because of its easy inoculability,
we usually supply it to those who wish to make therapeutic tests. One
naive experimenter, after working with this tumor, wrote that he
wished we would send him a better one, because although the tumor

PROCEEDINGS

he had previously used was curable by his treatment, this no. 180 was
worthless because it was too resistant. The value of his cancer cure is
evident.

As to the number of animals to be employed : we are in the habit of
using small series of animals over a long period of time; instead, for
instance, of using three hundred animals for an experiment, and one to
three hundred as controls we use two sets of twenty-five animals each,
and then repeat the experiment several times, employing a variety of
types of tumor, both carcinoma and sarcoma. We thus get long
series which can be compared with each other. A few months ago
there was published a statement concerning certain effects of x-ray, the
conclusions being based on the fact that in twenty or twenty-five ani-
mals, treated in one way, there were no tumor takes, while in the
control animals there were 80 to 90 per cent of takes. It was, there-
fore, assumed that an immunity was proved. All it was necessary to
do to show that the point was not proved was to open one of our record
books of the tumor this experimenter was using and find a page on
which were listed twenty-four animals which had been inoculated with
the same tumor on the same day, without a single take, while the next
lot of twenty-four under the same conditions showed 60 to 70 per
cent of takes. The only way to avoid these errors is to repeat the
experiment time after time, not necessarily with a large number of
animals, and in this way eliminate random fluctuations of which we
know nothing, but which are due to the fact, too often forgotten, that
an animal is not a test-tube, and that the tumor graft, also, varies in
its biological activities.

As to the question whether or not we can stimulate or retard the
growth rate of a tumor, the evidence which we have accumulated
proves that it is impossible to draw any accurate conclusions as to
rate of growth. Some 3000 grafts of tumor no. 180 were planted in
healthy mice from one dealer, allowed to grow for three weeks, and then
excised and weighed. It was assumed that the average weight of
these tumors would furnish a useful standard, yet in the next series of
200 tumors the average weight was found to be double that of the
previous 3000. When precautions are taken to have the site of inocu-
lation, the strain, the weight and the age of the animals, and the
method of feeding the same in all cases, and yet great fluctuations
occur, we are justified in saying that no conclusions as to growth rate
can be drawn. As . the result of experiments in which only a small
number of animals was used and no controls were kept, some obser-
vers have claimed to have stimulated tumor growth by chemicals;
but while such claims may be correct, they are absolutely unproved
because there are no available measurements by which growth rate
can be determined.

These are some of the complications of this tumor work, and the
only way to eliminate them is to know the biological qualities of the
tumor we are using, and to adjust the numbers of animals used to the
conditions of the experiment.

100

PROCEEDINGS

Dr. Stone: It seems to me that the purpose of this presentation is an
exceedingly good one. It brings to my mind the experience we are
repeatedly having at the Memorial Hospital in regard to radium.
Men go to Pittsburgh, for example, and make a contract for radium.
Then they come to the Memorial Hospital to learn how to use it; and
after a few days they return to Pittsburgh and get their radium. The
conclusions drawn from their work certainly can be no more accurate
than those which are drawn from the observation of the effects of
therapeutic measures on a few animal tumors.

6. The Influence of Certain Diets upon Tumor Susceptibility
and Growth in Albino Rats

Dr. K. Sugiura and Dr. Stanley R. Benedict (New York) :

SUMMARY

Curves constructed so as to show the linear and the percentage
growth of the fetus and of the Flexner-Jobling rat carcinoma in nor-
mally fed animals indicate that there is quite close agreement in growth
between the two (the carcinoma and the fetus).

The experiments reported were designed to determine whether
general differences in diet influence tumor growth independent of
special deficiencies, and to study the effect of certain specific defi-
ciencies in diet upon tumor susceptibility and growth. A complete
diet, composed of banana, 83 per cent, yeast, 0.5 per cent, and protein-
free milk, 0.5 per cent, was found to yield the same percentage of suc-
cessful inoculations and the same rate of growth as was secured with a
diet of wheat bread and whole milk. The elimination of certain
accessory factors in the banana diet was found not to influence tumor
growth except when the diet was restricted wholly to bananas. In
this latter case, the percentage of successful inoculations was the same
as in control animals, but the rate of growth of the tumor was markedly
retarded. Such dwarfed tumors showed no histological changes from
the normal, and resumed normal growth when the animal was placed
upon a normal diet.

discussion

Dr. Wells: Dr. Benedict's paper giving the curves of fetus and tumor
growths recalls a discussion which I heard when a paper from Ehrlich's
laboratory was presented. Some German with a mathematical turn
of mind had figured the rate at which tumor cells grow. At the end
of a certain number of days, he said, one would have enough tumor to
inoculate ten more animals, and at the end of three years the total
tumor tissue would have formed a cube so large that it would take a
ray of light 105 years to pass one side of it. Some of the biologists
present said there was nothing remarkable in that, because fetal tissues
had been demonstrated to have a fully equal capacity for growth.

PROCEEDINGS

101

Dr. Wood: I did not know when I spoke of tumor growth rates that
Dr. Benedict was going to present anything which turned upon the
details of growth rate, and I still hold the same views that I expressed
a few minutes ago. My opinion of Dr. Benedict's experiments is that
they may be right and they may be wrong^ The only way to prove
them right is to take the two significant series and repeat them on a
group of twenty-five rats, some ten times. This will eliminate random
fluctuations. Then if the averages show a striking difference between
the dieted and the control groups the fact may be accepted that diet
influences the type of tumor used, though not any other tumor, and
especially not a primary tumor. I can take from my own record
books pages that show very much diminished tumor growth rates
running over a period of two or three months, followed by a period of
very abundant growth, although the animals were untreated. I
cannot say, therefore, that Dr. Benedict is right or is wrong and the
only way to get proof either way is to repeat the experiment under a
great variety of conditions. It is the same problem as the question
whether radium does or does not stimulate a tumor. Many workers
believe that threshold doses of radium do stimulate but I have been
unable to prove it.

7. The Effect of Combined Heat and Radiation upon
Transplanted Animal Tumors

Dr. George L. Rohdenburg and Dr. Frederick Prime (New York) :

SUMMARY

The investigations of Loeb, Stevenson, and others have shown that
neoplastic cells are killed in relatively short periods by comparatively
low degrees of heat. In the present experiments mice and rat tumors
were cut into small fragments suitable for inoculation, exposed to
varying degrees of heat for varying periods while in Ringer's solution
in a water-bath, and then inoculated into animals, the lethal effect of
the manipulation being indicated by the percentages of takes in the
inoculated animals. In this fashion exposures to 40, 41, 42, 43, 44, 45,
and 46°C. were made for 15, 45, 75, 135, and 195 minutes. The lethal
effect of heat was manifested first with 40°C. at the end of 195 min-
utes, and with 41°C. at the end of 75 minutes; with 44° there were only
35 per cent of takes at the end of 45 minutes.

Bovie has shown that albumins which have been radiated coagulate
at much lower temperatures than those not radiated, and it was thought
it might be possible to apply this principle to the killing of the cancer
cell. Fragments of tumor prepared as were the fragments used in the
heat experiments were exposed to z-ray for periods varying from
10 to 30 minutes, 30 minutes representing three erythema doses.
Directly after radiation they were exposed to varying degrees of heat
after the method outlined in the previous paragraph and then inocu-

102

PROCEEDINGS

lated into animals. It was found that whereas heat alone at a given
temperature was not lethal, and x-ray in a given dosage was also not
lethal, the two when combined were lethal.

In another series of experiments the process was reversed, heat being
applied first, and radiation afterwards. The same effect was noted,
the sequence being apparently immaterial.

These results suggest a new method for treatment in human cases,
and experiments are now under way in which a practical application of
the principle is made on both human and spontaneous tumor animal
material.

DISCUSSION

Dr. Greenough: We have felt very strongly that the observation of
Dr. Bovie in regard to heat sensitization was one of possibly very great
clinical importance, but that further work should be done along experi-
mental lines before any serious attempt is made to apply it to living
patients. The point of the greatest importance is that a degree of
heat which under ordinary circumstances can be withstood perfectly
well by the individual cell results in the death of that cell if it is applied
after radiation. The temperatures which Dr. Rohdenburg has indi-
cated on the charts apparently are such that a definite effect can be
obtained in a certain small percentage of cases from the temperature
alone; to that extent, therefore, the experiment is not quite comparable
to Dr. Bovie's original observation; but I am very much interested in
the results, and I feel that without question this opens up a field for
clinical application in the actual treatment of cases.

Dr. Wood: I think everyone realizes now that we have reached a
point where we can say definitely whether or not radium and x-ray will
cure any malignant tumor. The only question is will the patient sur-
vive the dose. In other words, we have come to a point where it is a
question of operative mortality as compared to x-ray or radium mor-
tality. It is perfectly possible to obtain the absolute destruction of
cancer cells by a sufficient quantity of x-ray; and the same thing can be
done with radium. The only question is the practical application.
In my opinion, at the present time most internal tumors require so
much radiation that serious, very often fatal, damage to the normal
organs would inevitably be inflicted. For example, the lethal dose to a
carcinoma cell of a tumor 10 cm. below the skin would be between
sixteen to twenty erythema doses, if a highly filtered x-ray were used.
I have seen quite serious general disturbances to the intestinal tract
result from two erythema doses. The recent German literature, also,
contains reports of a number of deaths, some apparently from intes-
tinal obstruction, following heavy doses of x-ray. The problem is to
kill every cancer cell, although it is recognized that much benefit can
be obtained by the destruction of the main portion of the tumor, the
central areas, for example, which are poorly vascularized and easily

PROCEEDINGS

103

subject to necrosis from thrombosis of the capillaries of the tumor.
But the destruction of the peripheral portions of the tumor, which are
well vascularized, is quite another matter. Hence, a very large dosage
is required to kill absolutely all the tumor cells, either in the test-tube
or in the animal itself. Thus, although a point has now been reached
where we can say definitely that cancer can be cured by radiation,
with the dosage determined experimentally in this laboratory, the
question still is how we can also save the patient's life. With this
point kept clearly in mind Dr. Rohdenburg started his experiments.
The original experiments Dr. Prime and I began some four or five
years ago. The lethal death-points for cells were determined and are
standard, so there can be no difficulty in reproducing these results
anywhere if tumor 180 is used. An x-ray machine can be calibrated
biologically. My clinical experience shows that in a rapidly growing
carcinoma in man the death point is about the same as in 180. The
slow growing squamous-cell epitheliomata seem to have extra resis-
tance, owing to the ability of the cells to cornify and practically to
enter a resting stage. It is now accepted that the lymphosarcoma and
many of the bone tumors are more susceptible to x-ray than are the
very rapidly growing highly malignant carcinoma or sarcoma. The
maximum practical dose of x-ray or radium which the patient can
survive remains to be determined, but if it is remembered that there
is, after all, a surgical mortality and that untreated cancer is fatal, we
are justified in giving radiation pretty close to the limits. Most
radiation treatments are far beneath the killing dose for the cells of
most of the internal tumors. I do not believe that the induced con-
nective tissue reaction around the tumor cells kills them. It may
encapsulate them for a while, but usually the cell enclosed with scar
tissue finally begins to grow and destroys the patient. After having
given the maximum dose of x-ray which the skin of the patient can
stand, we might be able by thermo-electrical devices to apply heat
which does not seriously damage the skin and which produces death
of ^the cells. The morphology of cells killed by heat is practically the
same as that of cells killed by x-ray or radium, and I think that this
paper is important from a therapeutic point of view, as well as a scien-
tific one. It may be possible by some high frequency method to add
to the semi-lethal dose of x-rays a semi-lethal dose of heat without
causing the death of the patient and that, I think, these experiments
clearly show. The practical application will require long careful study
on human beings.

Dr. William Duane (Boston) : In regard to the practical application
of heat and x-rays, this was tried very extensively in Paris seven years
ago, and there was claimed a good deal for the combination. There
was one difference between that procedure and the one reported here,
the difference being that the two destructive agents were applied at the
same time. Keating-Hart warmed the human tumor tissues and at
the same time applied x-rays, and I understand that here heat was

104

PROCEEDINGS

applied either before or after radiation. It would seem probable that
the application of the two at one and the same time would produce a
greater effect. As regards the ability to destroy any tumor by radia-
tion, if an intense enough radiation is used almost any organic com-
pound can be destroyed. Water can be decomposed by the action of
the rays.

Dr. Rohdenburg: I would like to say that these experiments differ
from those previously reported in that lower degrees of heat were used
over a much longer period. In the recent German literature the com-
bination is reported to have been tried, but tried by combining diath-
ermy produced by seven amperes of current for a period of thirty
seconds. This dosage produces coagulation necrosis and a typical
burn. Our aim is to obviate this factor of burn. The tumors we have
treated with high frequency, using 25 milliamperes per square inch
of electrode surface and continuing the application for twenty minutes,
show scarcely any recognizable change in the tissue for a period of a
week or ten days. This dosage produces a temperature between the
electrodes of 40°C, or 104.8°F. which is very readily withstood by any
normal tissue, and many degrees below the temperatures produced by
Keating-Hart or the more recent investigators.

8. Interstitial Injections of an Active Deposit of Radium
Emanation in a Rat Carcinoma

Dr. Halsey J. Bagg (New York) :

SUMMARY

In this experiment a definite attempt was made to use an " active
deposit" of radium emanation as a local agent, employed in the form
of a solution, to control the growth of an experimental rat tumor —
the Flexner-Jobling rat carcinoma. The writer gave a preliminary
report of the results of treating twenty-one animal tumors.

Interstitial injections of a radio-active salt solution of radium emana-
tion were found to retard materially the growth of the tumors and in
some cases to cause their regression, resulting in characteristic cellular
changes in the tumor tissue. Definite characteristic, histological
radium changes were noted — frequent hyperchromatic and homog-
eneous nuclei, cellular hydrops, fairly extensive central necrosis, and a
terminal onset of fibrosis.

It was found that comparatively large doses of radium could be in-
jected into rat tumors without the escape of the radio-active solution
into the surrounding tissues. (A comparatively small intravenous or
subcutaneous injection of the same solution invariably results in severe
reactions.) The rapid decay of the radium when used in this form, in
addition to the walling-off effect of the tumor capsule, appears to be
sufficient to confine the radium action to a desired zone of tissue. This
was an encouraging observation from the standpoint of human therapy.

PROCEEDINGS

105

DISCUSSION

Dr. Duane: I think it would be well to emphasize the very great
danger of this method of using radium or radio-active substance. The
alpha rays are utilized, and since of the total activity of all the beta,
gamma, and alpha rays, about 90 per cent is due to the alpha radiation,
it is evident that the alpha rays produce relatively tremendous effects.
It is well to bear in mind, therefore, that this method of treating tumors
should be employed with very great care. Over 100 millicuries of
deposited activity would be a dangerous dose for any human being.

Dr. Benedict: I should like to ask if this is not simply a question of
the size of the tumor mass; that is, one might destroy a small tumor,
but affect only the center of a large one, making it necrotic.

Dr. Stone: It seems to me that both of these methods, the intra-
venous injection of the active deposit and the local infiltration of the
tumor, represent a type of experimental work which is most desirable,
but, as Dr. Duane has just said, we must use extreme caution in apply-
ing this practically. The general constitutional effects from intra-
venous injections are very marked. So far we do not know the con-
stitutional effects from the infiltration method, because it has not been
applied except in two or three instances, and then very cautiously.
It does seem, however, as if we might make use of the alpha rays in
this infiltration method, which, as Dr. Duane has said, comprises so
much of the energy of radium, and which apparently we cannot use
in any other way. There is another point in regard to the practical
application of this infiltration method. Personally I dislike to think
of radium used either by this method or by the surface application to
the extent that we get actual destruction or caustic effects. I conceive
of using radiation for a strictly biological effect, without causing abso-
lutely local destruction, for years prior to the use of radio-activity all
of us had experience with the use of the cautery, and we knew just how
much tissue we could destroy with that. With the use of radio-activ-
ity we do not know just how extensive the destruction is, how long it is
going to continue, or what the destructive effects upon the normal
tissues in the neighborhood of the tumor are. The possibilities of
destruction are so great that we must use extreme caution. In making
a practical application of this infiltration method I think we must
begin with the very smallest dose so as to get the so-called biological
effect, and not a local caustic effect on the tumor itself.

Dr. William B. Coley (New York): The work of Dr. Bagg is ex-
tremely interesting and valuable, but I believe that the cautions given
by Dr. Stone and Dr. Duane are timely. At the Memorial Hospital
Dr. Janeway has used the method in a considerable number of cases in
human beings, beginning with small doses and increasing to fairly good
sized ones; but so far none of the results gives us reason to believe that

106

PROCEEDINGS

it is likely to be of permanent value as a therapeutic measure. I
believe, as Dr. Bagg points out, that the cases in which good results
are obtained are those in which the tumor is localized by the surround-
ing tissues, and if that is so, why can not one use the bare tubes which
will remain localized, and not get into the tissues and damage the
kidney, liver, etc.? I believe with Dr. Wood that we can always give
enough radium to kill the cancer cells, but not in most deep seated
cancers without killing or injuring the patient.

Dr. Bagg: Dr. Benedict has brought up the subject of the relation of
tumor size to the subsequent reaction to the treatment. I have found
that in most cases the central portions of the treated tumors were
necrotic, and Dr. Ewing, who has seen my sections, has called my
attention to the peculiar manner in which the radio-active solutions
have diffused to different parts of the tumor, causing more extensive
reactions in certain areas of the tumor than in others. I have tried in
some recent work to infiltrate the tumors from within, by first placing
the hypodermic needle in the center of the mass and then moving it to
different parts of the periphery of the tumor, and injecting a small
amount of the solution in each place. In regard to the apparent
increase in size of some of the treated tumors, I would say that this is
no doubt due to the production of an edematous condition and the
accumulation of fluids in the center of the tumor, in which case the
ring of tumor tissue was crowded to the periphery, while the actual bulk
of the tumor tissue was no greater than the amount present at the
beginning of the experiment.

9. Factors Underlying Susceptibility to a Transplantable

Tumor in Mice

Dr. C. C. Little:

SUMMARY

A sarcoma, J. W. B., which originated in an inbred race of Japanese
waltzing mice grows upon transplantation in 100 per cent of the ani-
mals of that race inoculated. It fails to grow progressively in more
than 99.5 per cent of the common non-waltzing stock mice inoculated.

Beginning two weeks after inoculation each animal is observed at
weekly intervals, and the presence or absence of growth is noted.
When a growth is present the animal is recorded as + ; when no growth
is found, it is recorded as — . Two groups of animals have been
observed. They are (1) ordinary non-waltzing stock mice series (N.) ;
(2) back-cross hybrids (B. C.) produced by crossing an Fl generation
hybrid between Japanese waltzing and common non-waltzing mice
back with the common non-waltzing parent race.

The present series of experiments deals with growth of the tumor
from the second to the sixth week after inoculation, inclusive. Many

PROCEEDINGS

107

of the animals showing a growth at the sixth week observation showed
eventually regression and disappearance of the tumor. The factors
studied are, therefore, those allowing the initial six weeks' growth of
the tumor. This growth may or may not be continued later, depend-
ing upon the hereditary constitution of the mice used. The animals
inoculated were divided by age into ten groups, respectively, 2, 4, 6,
8, 10, 12, 14, 16, 18, and 20 or more days old at inoculation. For each
of these groups the percentage showing growth at 2, 3, 4, 5, and 6
weeks after inoculation was determined. The results of all groups
according to weeks after inoculation were as follows:

WEEKS

N., PER CENT +

B. C, PER CENT +

16.1

27.1

13.9

21.0

10.7

12.4

7.8

13 6

5.3

13.1

It will be noted that the percentage of animals showing growth in
the N. series steadily decreases, while in the B. C. series, there is an
initial decrease followed by a rise to a level at the fourth week at or
near 13 per cent. The total per cent for the N. series is 11.12 =±=46
and for the B. C. series is 17.54 ±83. The difference is 6.42 ±95 and
is, therefore, significant, being 6.7 times its probable error.

Studied by age groups with the sexes combined, there is seen to be a
distinct difference between the two groups.

AGE AT INOCULATION

N., PER CENT +

B. C, PER CENT +

days

2 to 10

12.87±0.6

13.77±1.05

12 to 20+

9.45±0.6

21.58=1=1.28

While in the N. series the lower age group shows the most growth,
exactly the opposite is true of the B. C. series in which the higher age
group shows a significantly higher percentage of growths.

When the sexes are studied separately the males of both series and
of both age groups are apparently not significantly different from each
other. The females, on the other hand, show a significantly lower
percentage of growths in the higher age group of the N. series and a
significantly higher percentage of growths in the higher age group of the
B. C. series. The higher age group of both sexes is more mature and
further differentiated than is the lower age group. Within the higher
age groups many of the females become sexually mature during the
period of observation. This gives to them an additional chance for
differentiation and the assumption of sex limited or other biological

108

PROCEEDINGS

characters dependent upon their hereditary make-up and upon the
degree of differentiation which their tissues have reached.

AGE AT
INOCULATION

B.C.

days

2 to 10
12 to 20+

15.70±1.64
10.30±1.20

19.46±1.31
9. 39 ±0.87

14.51±1.51
15.38±1.55

12.12d=1.76
25.74±2.07

What we are observing, therefore, is a racial difference. In the N.
series a steady decrease in percentage of growths, especially marked in
the females, is found. In the B. C. series there is at first a decrease
followed by a rise, a level of about 13 per cent growth being reached.
This level is due for the most part to the presence in the B. C. genera-
tion of animals which will show permanent progressive growth of the
tumor. The results of both series coincide with and amplify the
explanation of the hereditary nature of susceptibility to the J. W. B.
tumor, advanced by Tyzzer and the writer in 1916.

A STUDY OF A LIPOMYXOSARCOMA WITH COM-
MENTS UPON THE ORIGIN OF THE FAT CELL

VICTOR C. JACOBSON

From the Pathological Laboratory of the Peter Bent Brigham Hospital and the
Department of Pathology, Harvard Medical School, Boston

Received for publication April 19, 1921

Neoplasms of mesenchymal origin are frequently the cause
of much controversy owing to a lack of differentiation into
tissue which would permit their being placed in a definite
organoid category. The term "liposarcoma" has been used
to denote malignant change in the connective tissue of a lipoma,
itself one of the most benign of tumors, and also to accent the
fact that the growing fat cells have invaded the surrounding
tissues or metastasized. A tumor which presents the accepted
criteria of malignancy, derived from fat cells which are atypical
in size, shape, and staining characteristics, with many mitotic
figures and invasion of septa and capsule, is of the greatest
rarity, and a study of such a growth may throw light upon the
origin of the fat cell, regarding which there is much conflicting
opinion. The writer has lately had the opportunity to examine
a tumor of that type, removed from a patient in the Surgical
Out-Door Department of the Peter Bent Brigham Hospital,
and with the kind permission of Dr. Harvey Cushing a brief
summary of the case is given.

The patient, a Russian Jew, came to the hospital on November 18,
1920, complaining of a growth in his right thigh. His past history
is unimportant except in so far as it concerns the tumor. In 1917
he noticed a firm swelling the size of a walnut beneath the skin on the
mesial surface of the right thigh. It caused no pain. A few months
later it was noticeably larger, and massage was instituted for a short
time. It continued to grow slowly and became a little painful, particu-
larly in cold weather. He was, however, able to work in his store,
and he walked into the hospital.

109

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 2

110

VICTOR C. JACOBSON

Physical examination was negative except for the local findings.
On the inner surface of the upper third of the right thigh, beneath the
skin and extending deep into the muscles, was felt a somewhat movable,
irregularly rounded growth, about 18 x 6 cm. in size, and not tender.
There were no palpable nodes. On November 19, under novocaine
anesthesia, the growth was removed by Dr. A. H. Brewster. It seemed
to have three distinct capsules, but there was no evidence of invasion
of the surrounding muscles. The outer two capsules were thick and
fibrous, and stripped away readily from the more delicate layer covering
the tumor. Most of the blood-vessels entered through a sort of pedicle
on the deeper portion of the tumor. The dead space was closed with
catgut sutures and the skin with silk. On December 21 the wound
had entirely healed. On February 3, 1921, the wound was in good
condition, with no evidence of local recurrence or metastases, and
the patient felt perfectly well.

Pathological examination. The specimen consists of an irregular mass
measuring 17 x 8 x 5 cm., and weighing 510 grams. It is surrounded
by a delicate transparent capsule beneath which ramify a moderate
number of dilated blood-vessels. The larger vessels converge upon a
narrow zone on one side which probably represents the pedicle where
the vessels were clamped off. It is uniformly firm, more so in slightly
depressed areas between the bulging large lobules into which it seems
to be divided. A longitudinal incision shows it to consist of two main
types of tissue. About three-fourths of the tumor is pale, rather pink,
and opaque, with the consistence of a very cellular tumor. The mass
is divided into more or less definite lobules by fibrous septa of varying
thickness. A broad fibrous band apparently invaded by streaks of
tumor separates it into two main parts, which are in turn subdivided
into smaller lobules by narrow vascular septa. More or less well
circumscribed areas suggesting mucoid degeneration are present in the
middle portion, irregular wells of mucilaginous fluid bathing the surface
like honey in a comb. The fluid gives the characteristic reactions of
true mucin.

Microscopical examination. Tissue was fixed in 10 per cent neutral
formalin and in Zenker's fluid. Frozen sections were stained with
Scharlach R and hematoxylin, other preparations with methylene blue
and eosin, phosphotungstic acid-hematoxylin, aniline blue-acid fuchsin,
osmic acid, Mayer's mucicarmine and mucihematein, and Verhoeff's
elastic tissue stain. A study of the sections gives the following facts.

The tumor is surrounded by a thin compact layer of connective tis-
sue in which are a moderate number of elastic fibers. The capsule

STUDY OF A LIPOMYXOSARCOMA

111

sends prolongations into the tumor dividing it into smaller units.
Elastic fibers and thin-walled arteries and veins accompany the septa
and many fine capillaries are present in the stroma, which consists of
a delicate collagenous reticulum in which the tumor cells lie, and which
appears to arise for the most part from the blood-vessels. The tumor
is very cellular except in small areas which have undergone mucoid
change, where the tissue suggests that of the umbilical cord. In shape
the cells vary from medium sized spindle cells through intermediate
forms to the more or less round outline of mature fat cells. The nucleus
follows the general shape of the cell and has a densely staining nucleolus.
Practically all the cells contain refractive fat droplets which stain with
Scharlach R and reduce osmic acid. In the spindle cells the droplets
are, as a rule small, almost always smaller than the nucleus, and tend
to localize in the vicinity of the nucleus. These spindle cells often
have delicate wavy processes of some length. In the intermediate,
polyhedral forms the fat drops become larger and coalesce, the nucleus
stains more deeply, and mitotic figures are seen in large numbers. The
• dividing cells have a large fat content. The more mature round cells
are distended with a single droplet of fat which occupies practically the
entire cytoplasmic space and flattens the nucleus, giving the typical
signet-ring appearance. Many of these cells have ruptured, and the
fat has escaped into the interstices, where it has formed little pools.
Invasion of the capsule and septa by the tumor cells has occurred. No
intravascular growth is found, however. Scattered through the stroma
are also many mast cells, none of which contain oxidase.

Sections from the areas of "mucoid degeneration" show an abundance
of loose wide reticulum containing mucin. In this are present a moder-
ate number of tumor cells of all types with large and small fat droplets.
With specific stains mucin can be demonstrated in most of the cells
in the form of a fine or coarse granular deposit in the cytoplasm, clus-
tered about the fat droplets. Mucin is present in small amount in the
cells of various other parts of the tumor but is most abundant where it
exists in macroscopic quantities.

DISCUSSION

A lipoma is a very common and benign tumor composed of
mature fat cells. A liposarcoma is much less frequently met
with and in a large percentage of cases has shown areas of
myxomatous degeneration. Knox (1) in 1919 collected twenty-

112

VICTOR C. JACOBSON

five cases and added two more. The names lipomyxosarcoma,
myxoliposarcoma, liposarcoma myxomatoses, and myxosarcoma
lipomatodes have been used by various writers, depending
upon which tissue was in excess, the fatty or mucoid, or whether
the writer regarded the mucous tissue as a product of degen-
eration of the fatty tissue or vice versa. "Sarcoma" denoted
invasion, recurrence, or metastasis. Fibrous tissue, cartilage,
and bone have also been found in these neoplasms, making
them rather mixed tumors, although with all elements derived
from the same germ layer. Hirsch and Wells (2) studied a
retroperitoneal tumor of 69 pounds, parts of which were pure
lipoma, others fibrosarcoma. In none of the cases described
have mitotic figures been found in the fat cells, although evi-
dence of direct division has been noted. The most frequent
sites of these tumors are the retroperitoneal tissues, the muscu-
lar fasciae of the thigh and leg, the mesentery, the cheek, and
the breast.

The tumor forming the basis of this paper has several unique
features. The bulk of it is composed of cells in practically all
of which is much fat. Most of these cells are of the embryonal
type, being spindle or stellate with delicate fibrillary processes.
Many transitional forms are present, up to and including the
large round signet-ring mature fat cell. Many cells are in mitosis
and these also contain fat droplets. About one-fourth of the
tumor is bathed in mucin, and the cells in such places contain
both fat and mucin.

In offering an explanation for the presence of mucin and fat
in the same cell, we must go back and consider the embryo-
logical development of the connective tissues. The connective
tissues which include fat and mucous tissue are derived from
the mesenchyme. Todd and Bowman (3), in 1845, were among
the first to suggest that adipose and areolar tissues are distinct
and independent. Kolliker, in 1856, described the original fat
cells as glands, such being found in the axillae and mesentery,
and around the kidneys. In 1870 Toldt's (4) investigations
led him to the conclusion that the 1 'glands" of Kolliker are the
source of fat, and he called them "fat organs." Flemming

Fig. 1. Photograph of the Original Tumor, Exterior

114

VICTOR C. JACOBSOX

(5) in 1876, in a comprehensive study, advanced the opinion
that the ancestral tissue of fat is a fibrillar connective tissue.
Hammar (6) also held this view.

In 1894, Borden (7), in a comparative study of the fat cell,
stated that (i) in the lower vertebrates the fat cell is developed
from one form of cell arising from special centers; (ii) in the
higher vertebrates the fat cell is developed from two forms of
cells differing greatly in size and shape; and (hi) one of the
original tat cell forms in the higher vertebrates is homologous
with the special center forms of the lower vertebrates. This
cell is gland-like and in no way resembles the connective-tissue
cell; the other form, while closely resembling the connective-
tissue cell, is most probably a special cell derived from the
special center forms by cell division and "migration metastasis."
Multiplication is by indirect division and the cells never turn
into connective tissue cells.

Bell (8), working with material from a calf fetus and from
steers, described a peculiar open-meshed "pre-adipose" tissue
which forms well defined lobules before true fat cells appear.
This "pre-adipose" tissue consists of loosely arranged cells
with two or more coarse processes. The branched cells may
contain tat droplets a long time before they assume the rounded
form. Bell supports Flemming in the belief that the pre-
adipose tissue is clearly a fibrillar connect ive tissue. He also
described how the cells nearest the blood-vessels filled with
fat first, the processes becoming absorbed. Altmann's granules
were observed when the cells were yet branched and before the
fiisi fat droplets had appeared.

in Lewis and Stohr's Textbook of Histology (9) it is stated
that in the four-month embryo the fat cells are quite like the
surrounding fibroblasts, being fusiform or stellace and con-
taining vacuoles or droplets. Mallory (10) emphatically says
that the fat cell is a perfectly definite type of cell formed by
differentiation from a mesenchymal cell, that it is not a fibro-
blast, does not arise from one, and, when il undergoes atrophy,
does not t urn into one.

STUDY OF A LIPOMYXOSARCOMA

115

The earlier writers, in speaking of the connective tissue cell,
probably referred to what is now termed the fibroblast. Conse-
quently, at the present time the opinion is strongly held by
many observers that the fat cell is essentially a modified fibro-
blast, whereas others assert that it cannot be so considered.
I believe that the demonstration in this tumor of mucin in
the cells which otherwise appear as fat cells supplies an argu-
ment in favor of the former idea.

Mucin is a slimy substance produced by certain epithelial
(beaker) cells and by the fibroblast under special conditions.
We are concerned here with its production by the latter cell.
The umbilical cord furnishes the classical example of mucous
tissue. The mucin is said to be secreted by the fibroblasts
without the formation of special granules or vacuoles (Lewis
and Stohr) . Cohnheim restricted the term mucin to those sub-
stances which are excreted by epithelial cells, and reserved
the name mucoid for closely allied bodies which occur in various
organs of the body, such as the vitreous humor, cornea, umbili-
cal cord, and tendons.

Mucin has certain characteristic chemical and physical
reactions which permit its identification, so that the term
mucoid seems unnecessary except for describing gross physical
qualities of the fluid. It is not produced normally in the adult,
but in tumors of mesenchymal origin it is found not infre-
quently, as in myxoma, fibroma, chondroma and osteoma,
and in the more rapidly growing myxosarcoma. Ewing (11)
states that mucous tissue "is genetically related to fat tissue
into which it is extensively transformed during normal growth.
. . . . It does not appear that true myxomas ever tend
to differentiate into fibroma or lipoma." He also asserts that
the spindle and star cells in the mucous matrix may contain
hydropic or fatty droplets.

The question naturally arises, is the substance in the tumor
cells which stains with Mayer's mucin stains really mucin?
The tissue in which these cells are found in greatest numbers
is bathed in a slimy fluid which gives the accepted tests for
mucin and likewise stains with mucicarmine and mucihematein.

116

VICTOR C. JACOBSON

It is, of course, impossible to be certain that the intracellular
substance is mucin, yet it stains "specifically" and any other
interpretation would seem illogical. The cells otherwise have
the appearance of young fat cells such as are found in other
parts of the tumor.

It might seem unwise to attempt to settle the origin of a cell
solely by study of a tumor derived from that type of cell, particu-
larly in view of the frequency of metaplasia in pathological
experience. But metaplasia apparently has a very narrow scope,
its best illustrations being in epithelial reactions, and applica-
tion of the principle to the connective tissues has led to great
differences in interpretation. To explain the secretion of mucin
by the cells of this tumor as an example of metaplasia does
not seem as rational as to assume that the cells which are
essentially fat cells take on the characteristics of a less differ-
entiated tissue, a noteworthy deviation from the normal being
that the mucin is stainable in the cells.

These observations, it seems to me, indicate that the fat
cell and the fibroblast (considering the mucous connective-
tissue cell as a modified fibroblast) are very closely related and
give support to the hypothesis that the fat cell is derived from
the fibroblast.

REFERENCES

(1) Knox: Proc. New York Path. Soc, 1919, xix, 90.

(2) Hirsch and Wells: Am. J. M. Sc., 1920, clix, 359.

(3) Todd and Bowman: Physiol. Anat., 1845, i, 90.

(4) Toldt: Quoted by E. A. Schaafer, Textbook of Microscopic Anatomy,

New York, 1912, p. 130 (Quain's Anatomy, vol. ii, pt. 1).

(5) Flemming: Arch. f. mikr. Anat., 1876, xii, 434.

(6) Hamhab: Arch. f. mikr. Anat., 1895, xlv, 512.

(7) Borden: New York M. J., 1894, lix, 225.

(8) Bell: Am. J. Anat., 1909, ix, 401.

(9) Lewis an<1 Stohr: Textbook of Histology, 2d ed., Philadelphia, 1917, p. 62.

(10) Mallory: Principles of Pathologic Histology, Philadelphia, 1914, p. 84.

(11) Ewino: Neoplastic Diseases, Philadelphia, 1919, pp. 166, 167.

PLATE

117

PLATE 1

Fig. 3. Drawing of microscopical field, high power. Frozen section stained
with Scharlach R and hematoxylin. The distribution of fat in the cells and
interstices is shown. In the center are two cells in mitosis and containing fat
droplets.

Fig. 4. Drawings of four cells in a section stained with Scharlach R and
hematoxylin, showing the transition from spindle cell to the mature signet -ring
fat cell.

Fig. 5. Drawings of three cells in a section stained with Mayer's mucicarmine,
from an area of myxomatous tissue. The finely granular mucin is seen in the
cytoplasm between the fat droplets.

118

PARABIOSIS AND TUMOR GROWTH

ISIDOR KROSS, M.D.
From Columbia University, Institute of Cancer Research, F. C. Wood, Director

Received for publication June 15, 1921

The subject of tumor immunity is of considerable importance,
and any approach that promises to throw light upon this as yet
unsolved problem should be thoroughly investigated. The effect
of parabiosis upon tumor growth has been studied by various
investigators, but their results and conclusions have been far from
uniform. In an attempt to clear up this matter, the experi-
ments here reported were undertaken.

Sauerbruch and Heyde (1), in a series of experiments with
parabiotic animals, noticed that the death of one animal was
always followed by the death of the other in a few hours unless
the live animal was cut away from its dead partner. This,
they maintained, is probably due to the absorption of cadaver
toxins from the dead animal. They also noticed that when
iodine solution was injected into one animal, the urine of the
other gave a positive iodine reaction within forty-five minutes.
That corpuscular elements also could pass from the one to the
other, they proved by injecting a culture of anthrax bacilli into the
right animal and recovering the characteristic organisms from
the heart's blood of the left. From their anatomic studies, the
authors conclude that the operative wound uniting the two ani-
mals heals exactly as does that in a single animal, except that
the reaction is more intense, owing probably to "the foreign
body" reaction.

Friedberger and Nassetti (2) studying the antibody formation
in parabiosis found that agglutinins formed as a result of the
injection of typhoid bacilli or Vibrio elbensis into one animal
could be recovered from the non-injected one. They also demon-
strated that one animal could be immunized passively by the

121

122

ISIDOR KROSS

injection of these organisms into the other. Morpurgo (3), in
a series of most interesting experiments, showed convincingly
that after a bilateral nephrectomy was performed upon one rat,
the kidneys of its partner in parabiosis were able to compensate
for their loss and maintain life for both animals — in one instance
for forty days. In another case, he succeeded in keeping two
parabiotic rats alive for seventeen days after unilateral nephrec-
tomy in one animal, the two kidneys of the other having been
previously removed. Jehn (4) and Sauerbruch and Heyde
obtained similar results in rabbits.

The findings in all these investigations show how very intimate
is the biological relationship between the parabiotic animals.
Lambert (5) studied the influence of parabiosis upon the growth
of transplantable tumors in rats, and concluded that the
growth of mouse tumors in rats is definitely promoted by
parabiosis in that the percentage of successful inoculations is
increased, the rate of growth accelerated, and the duration of
active growth extended. Rous (6) in a series of experiments
with rats concluded that parabiosis has no effect upon the
growth of tumors. Albrecht and Hecht (7), however, showed
that parabiosis in mice exerted a distinct retarding influence
on the growth of tumors, and explained this phenomenon as
similar to the immunizing action of subcutaneous injections of
blood, embryo, or normal tissue.

In a previous study (8) the author showed that the intraperi-
toneal injection of blood from immune animals into susceptible
ones was without retarding influence upon the growth of tumors
in the latter. This paper is a continuation of the previous
study and is based on the following working hypothesis: We
assume that an immune animal possesses substances antagonistic
to tumor growth, or that it lacks those elements necessary to
tumor for its growth. Having proved that these substances
do not reside in the circulating blood, we further assume that
they exist somewhere else in the organism. By performing a
parabiosis between a refractory and a susceptible animal,
we bring these two organisms into very intimate biological
relationship. We have further assumed that it may be possible

PARABIOSIS AND TUMOR GROWTH

123

(1) for the " antibodies" in the immune animal to cross over into
the non-immune parabiotic partner and confer upon it the
property of antagonism to tumor growth; (2) for the substances
that make for tumor development in the susceptible animal to
cross over into the immune one and render that animal suscepti-
ble. To test these possibilities this study was undertaken.

In one part of these experiments, 44 pairs of animals were
united in parabiosis. Two distinct breeds of animals were
employed — August rats, which are susceptible to the growth of
Jensen rat sarcoma (J R S) and Marshall rats which are refrac-
tory to this tumor. Five days subsequent to the uniting of two
of these animals, the Marshall rat was inoculated with 0.003
gram of Jensen rat sarcoma. Of the 44 pairs, 3 died before
the tumor inoculation. The parabiotic life period of the others
varied from four to thirty-five days. Five pairs were lost leaving
36 pairs from which to draw conclusions. In only sixteen instan-
ces (44 per cent) could tumor formation be made out and at that
the majority of these growths were pinhead size. Of the control
animals, individual rats of the same breed receiving similar inocu-
lations of Jensen rat sarcoma, 32 out of 43 (74 per cent) showed
positive results. These tumors were larger, appeared earlier,
and lasted longer than those in the parabiotic series.

In chart 1 are shown the tracings of the growth of the Jensen
rat sarcoma in the Marshall rat. The gradual growth soon
reaches its peak and in a short period of time the tumor recedes
and disappears completely. In the parabiotic Marshall rats,
in practically all cases where the tumor does develop, it fails to
grow as well as in the controls and recedes much more rapidly.
In the first two instances on chart 2 it should be noted that the
rapid growth took place only after the August animal had died.
The obvious conclusions are (1) that the susceptible animal does
not transmit any of its characteristics to the immune one; (2)
that parabiosis reduces the rate, intensity, and duration of
growth of the Jensen rat sarcoma in the Marshall rat. This is
probably due to the ill health of the animal in parabiosis.

In a second series, tests were made to determine whether
protective substances are transmitted from the immune to the

124

ISIDOR KROSS

susceptible animal, in other words, whether passive immunity
can be developed in this manner. The susceptible Marshall
rats were inoculated with Flexner rat carcinoma (FRC). Nine

J RS INTO MARSHALL CONTROL

MAY 19.20 24

JUNE 1

•

^™

—

■

•

—

•

RS INTO

MARSHALL

PARABIOSIS

MAY 15

29 JUNE 1

SEPARATED

•

MAY 29

JUNE t

JULY 3

•

MAY 19

29 JUNE 1

•

JUNE 9

MAY 1

30 JUNE 4

^9_

•

• •

SEPARATED

10 CM

Charts 1 and 2

pairs were operated upon; six pairs survived the four day post-
operative period and were inoculated with the tumor. (Two
survived for a period of over two weeks, which allowed chartings

PARABIOSIS AND TUMOR GROWTH

125

enough to make a fair comparison with the control. Chart 3
shows that the growths in time of appearance, and rapidity of
development are only slightly behind those of the controls.

FRC INTO

MARSHALL

SEPT. 27-20

PARABIOSIS
MARSHALL AND AUGUST

OCT.I

S 9 13

—

— t OCT. 11.20

• •

•

f OCT. 19.20

•

«. . t 0CT.I2.20

— t OCT 14.20

FRC INTO

MARSHALL CONTROL

• • •

•

1 1

♦

Chaets 3 AND 4

CONCLUSIONS

From these experiments it is quite logical to conclude that
parabiosis does not increase the susceptibility of the immune
animal nor does it give immunity to the susceptible one, and that
whatever the substances are that make for growth in the suscepti-

126

ISEDOR KROSS

ble and for non-growth in the immune animal, these properties
are not carried over by the parabiotic union. The lesser size of
the tumor and the lesser rapidity of its growth in the susceptible
partner can be explained by the setback that the operation has
produced and by the interference with the normal activities of
the animal, and it is unnecessary to assume as do Albrecht and
Hecht that this phenomenon is one of partial artificial immunity
similar to that produced by the injection of homologous tissue,
since such immunity is known not to affect tumors after growth is
begun but only to prevent inoculation. While the number of
animals employed in these experiments is rather small, the results,
taken in conjunction with those of other observers in this field,
are sufficiently consistent to make it unnecessary to repeat them
in a larger series.

REFERENCES

(1) Sauerbruch and Heyde: Munchen. med. Wchnschr., 1908, lv, 153.

(2) Friedberger and Nassetti : Ztschr. f. Immunitatsforsch. u. exper. Therap.,

1909, ii, 509.

(3) Morpurgo: Verhandl. d. deutsch. path. Gesellsch., 1910, xiv, 259; 1909,

xiii, 150.

(4) Jehn: Ztschr. f. exper. Path. u. Therap., 1909, vi, 16.

(5) Lambert, R. A. : J. Exper. M., 1911, xiii, 257.

(6) Rous: J. Exper. M., 1909, xi, 810; Proc. Soc. Exper. Biol, and Med., 1909-10,

vii, 12.

(7) Albrecht and Hecht: Centralbl. f . allg. Path., 1909, xx, 1039.

(8) Krosb: J. Cancer Research, 1921, vi, 25.

THE PROTEIN CONTENT OF THE WHOLE BLOOD AND
PLASMA IN CANCER

RUTH C. THE IS

From the Huntington Fund for Cancer Research, Memorial Hospital, and the
Harriman Research Laboratory, Roosevelt Hospital, New York

Received for publication August 1, 1921

In 1919, Dr. William S. Stone and the writer (1) reported a
study of the chemical composition of the blood in cancer cases,
describing the results of determinations of the sugar and the
non-protein nitrogen constituents of the blood. It has been
a question for years whether the proteins of the blood are in-
creased or decreased in cancer. Only recently Robin (2) stated that
the serum protein is increased, but, like so many other investiga-
tors, he based his conclusions on the study of only a few bloods.
Therefore, it seemed advisable to investigate a sufficient number
of cases and compare the results with those obtained in other
diseases.

The nitrogen of the whole blood, serum, or plasma was deter-
mined by the Kjeldahl method, and the hemoglobin with Dare's
hemoglobinometer. Cancer bloods were obtained from the
wards of the Memorial Hospital, and the others from the Roose-
velt Hospital.

Forty-three bloods were examined. Tables 1 and 2 show
the results obtained on cancer bloods, the only difference being
that table 1 refers to serum and table 2 to plasma. Table 3
contains the results on bloods from other pathological conditions.

The protein content of the whole blood depends to such a
large extent upon the amount of hemoglobin present that we
should expect a large variation. Thus, in table 1 whole blood
protein varies from 12 to 19 per cent, with an average of 16.4
per cent.

127

128

RUTH C. THEIS

TABLE 1

Protein of whole blood and serum in cancer

NUMBER

DIAGNOSIS

WHOLE
BLOOD
PROTEIN*

SERUM
PROTEIN

DARE
HEMOGLOBIN

J)6T CCTlt

per CCTlt

per cent

Papilloma of bladder

12.1

7.7

Carcinoma of bladder

17.6

6.6

Carcinoma of bladder

16.2

6.0

Carcinoma of prostate

i a r\

19.0

7.3

Carcinoma of prostate

19.1

6.7

Carcinoma of prostate

12.1

6.9

Carcinoma of prostate

16.2

8*2

Carcinoma of parotid gland

17.8

6.9

Carcinoma of uterus

14.1

7.1

Carcinoma of rectum

18.1

6.4

Epithelioma of jaw

16.2

6.0

Epithelioma of tonsil

16.8

6.6

Neurosarcoma of arm

18.2

7.7

♦Nitrogen X 6.25 = protein.

TABLE 2

Protein of whole blood and plasma in cancer

NUMBER

DIAGNOSIS

WHOLE BLOOD
PROTEIN*

PLASMA
PROTEIN

DARE
HEMOGLOBIN

- per cent

per cent

Carcinoma of bladder

16.9

7.1

Carcinoma of bladder

13.2

5.3

Carcinoma of bladder

16.1

5.6

Carcinoma of bladder

18.8

7.1

Carcinoma of uterus

15.5

5.6

Carcinoma of breast

14.7

6.1

Carcinoma of breast

15.7

5.9

Carcinoma of rectum

14.4

8.1

Carcinoma of rectum

15.0

5.9

Carcinoma of liver

15.8

6.1

Carcinoma of larynx

14.7

6.7

Carcinoma of stomach

8.7

5.1

Teratoma of testis

18.2

6.7

Keloid of chest

19.3

7.4

*Nitrogen X 6.25 = protein.

In table 2 the protein varies from 13 to 19 per cent with an
average of 1G.3 per cent, if we except case no. 25, in which the
blood was very anemic with a protein value of only 8.7 per cent.

PROTEIN CONTENT OF BLOOD AND PLASMA IN CANCER 129

In table 3 the protein varies from 12 to 20 per cent with an
average of 17.6 per cent. Hemoglobin readings were, unfortu-
nately, not done in these cases but they would undoubtedly
have averaged higher than in the cancer cases and would account
for the slightly higher protein value of the whole blood.

If there is an increase or decrease in the amount of protein
in cancer blood, it would show in the serum or plasma. Here
we can see that the variation is very slight. Table 1 varies from
5.9 to 8.1 per cent with an average of 6.8 per cent; the patient

TABLE 3

Protein of whole blood and plasma in other diseases

NUMBER

DIAGNOSIS

WHOLE BLOOD
PROTEIN*

PLASMA PROTEIN

per cent

Hematoma of kidney-

19.8

7.2

Tuberculous glands

16.4

5.4

Colon bacillus infection

18.2

5.8

Paralysis agitans

19.6

7.9

Pleurisy

15.3

6.0

Ear infection

19.6

8.0

Retroversion of uterus

16.0

6.9

Urticaria

17.9

8.1

Possible adhesions

19.6

7.6

Neurasthenia

17.8

7.2

Neurasthenia

17.3

7.4

Pyelitis

18.4

8.0

Pyelitis

12.3

7.9

Pyelitis

20.0

5.9

Pyelitis

18.4

7.2

Pyelitis

16.3

8.0

♦Nitrogen X 6.25 = protein.

whose blood had the highest value died the day after the blood
was taken. The plasma protein values of table 2 varied from
5 to 8.1 per cent, with an average of 6.3 per cent. The plasma
values of other pathological conditions varied from 5.3 to 8
per cent, with an average of 7.0 per cent.

Loebner (3) reported results on 44 cases of cancer. She
employed the refractometric method, using serum obtained from
finger blood. Her figures varied from 4.2 per cent to 9 per cent
with an average of 7.3 per cent, which corresponds very closely

THE JOURNAL OP CANCER RESEARCH, VOL. VI, NO. 2

130

RUTH C. THEIS

with the average of this series. She considered 8.2 per cent the
average for normals since that is the average which Heudorfer
obtained in a study of five cases. This would indicate that
the protein of the serum is somewhat reduced in cancer.
If, however, we compare the average in cancer cases with the
average in cases of other pathological conditions, we can see
that there really is no difference in the amount of protein present.

CONCLUSION

Proteins of the blood plasma are neither decreased nor in-
creased in cancer cases as compared with other hospital patients.

REFERENCES

(1) Theis and Stone: Jour. Cancer Res., 1919, iv, 349.

(2) Robin: Bull, de PAcad. de meU, 1920, lxxxiv, 51.

(3) Loebner: Deutsch. Arch. f. klin. Med., 1918, cxxvi, 61.

PROBLEMS IN CANCER RESEARCH*

MONTROSE T. BURROWS

From the Department of Surgery, Washington University School of Medicine and
Research Laboratories of the Barnard Free Skin and Cancer Hospital,
Saint Louis, Missouri

Received for publication June 6, 1921

It is a pleasure to announce that a Department for Cancer
Research has again been established in St. Louis by the Barnard
Free Skin and Cancer Hospital. The material available is the
laboratories and what can be developed from the more general
biological side within them, and the clinical material.

The plan which is being used for the study of cancer divides
the work into several distinct groups. One is a general study of
the properties of the various cells of the body and the peculiarities
of the mechanisms of growth, division, differentiation, and func-
tion of these cells. A second is a close study of the clinical
course of the disease and the results of methods of treatment
and diagnosis. A third is an attempted analysis of isolated
facts now fully demonstrated, and their immediate relation to
clinical and pathological data at hand.

For the immediate time, one problem which seems to be
most pressing is the improvement of methods of diagnosis,
especially the diagnosis of internal cancer. Early diagnosis of
cancer of the breast has prolonged the life of many patients for
at least ten or twenty years. The same is true for other super-
ficial growths. Barring a few exceptional cases, methods for the
diagnosis of internal cancer have not reached the same degree of
perfection, as, for instance, cancers of the intestinal canal. If
by chance a cancer of the intestine causes early obstruction, a
diagnosis may be made and the treatment may be hopeful. How
few cancers of this region give, however, this early symptom!

1 Presented before the American Association for Cancer Research, Cleveland,
Ohio, March 24, 1921.

131

132

MONTROSE T. BURROWS

Another of the immediate problems is the betterment of
methods of treatment now available. The third and most im-
portant of all is the etiology of this disease. It is most impor-
tant, because through its elucidation and even in its study all
other problems must become simplified. Through it alone,
barring some peculiar accident, can the hopes for ultimate
success in treatment come. The greatest effort in cancer must
be exerted, therefore, in this direction. .

What is known about cancer today, aside from clinical con-
siderations, is little more than was known many years ago. This
is appalling but true, and the question arises: Is it not due to
the fact that the general notions of the cell, its structure and
properties have not materially advanced during this time? The
problem of the etiology of cancer when resolved into terms of the
cell may be quoted thus : Is the cancer cell a cell different from
normal cells, or is it a normal cell suffering a continuous external
growth stimulus? Since the application of the theories of cellu-
lar growth now in vogue have failed to answer this question,
there remain open but two channels of attack; the hit and miss
method, and a further probing into the nature of the mechanisms
of cellular growth, division, differentiation, and function.

Cancer, no matter how it is taken, represents a break in the
normal growth-regulating conditions of the body. Without a
knowledge of the normal growth-regulating mechanism, it is to
be expected that the etiology of cancer will remain obscure, unless
it is proved that the disease is the result of some definite foreign
agent or stimulus — parasite, or otherwise — which may be isolated.

It is known that the body is composed of cells. The minute
metabolism of these cells, the mechanism of their formation, even
the simplest constituents of their life are, however, based only on
theories. The leucocytes are compared to the amoeba. One
talks glibly of cellular metabolism, when he probably means
body metabolism. Is this so-called cellular metabolism the
work of the cell or is it the work of the organ or the body as a
whole? In beginning a few years ago the study of the cells in tissue
culture, it was of interest to note that the leucocytes, both the va-
rious mononuclear as well as the polymorphonuclear types, show

PROBLEMS IN CANCER RESEARCH

133

no ability to grow and divide like other cells. Their movements
are passive reactions to external conditions. A connective tissue
cell or a heart muscle cell suffering from the effects of certain
toxic poisons can be made to resemble the small and large mono-
nuclear cells of the inflammatory exudate. These latter cells
have lost, also, all ability for further growth under the condi-
tions in which they had previously grown or under any other
conditions it has been possible to impose upon them. In actively
growing culture many of these fixed tissue cells, unaffected by
foreign poisons may also show such changes after they have
been preyed upon by actively growing cells.

The polymorphonuclear leucocyte is a cell differentiated in a
special locality. Lymphocytes undoubtedly are found in the
same locality as well as in lymph-nodes. Whether the large
number of mononuclear cells of the chronic inflammatory area or
about cancers have this origin is still a very much open question.
In fact, the work of Maximow has not been confirmed. (See
Stockard (1).) That these cells contain enzymes and bacterio-
lytic substances which are liberated when they break down is,
again, well established. What is not known are the conditions
under which these cells live in the areas about a cancer and
whether they do liberate these substances under ordinary condi-
tions there. It seems evident, therefore, that they might as
well be considered the result of active growth or the presence
of toxic substances, as to be thought of as antagonists to these
conditions. In chronic inflammations, it is generally where the
bacteria are most active that they are most prevalent. In the
tissue culture it is the actively growing tissue cells that show the
true antagonistic action to growing bacteria. The lymphocytic
cells and the rounded-off connective tissue cells intermingle in
the cultures with the bacteria. The connective tissue cells,
the heart muscle cells, and other actively growing cells cease
to grow when they come within the zone of bacterial action.
The bacteria stand them off as they also stand off the bacteria (2).

This absence of an absolute knowledge of many of the most
important functions of body cells must in itself make it very
unsafe, therefore, to draw any far reaching functional conclusions

134

MONTROSE T. BURROWS

from morphological data. It makes it absolutely necessary to
draw one 's conclusions from the clinical facts alone. This applies
not alone to treatment, but also to other experimental results.
The laboratory in Saint Louis is well situated in this regard in
that it is a definite part of the clinic, and also allows the broader
type of biological research which in the light of present knowledge
must form an intimate part of the cancer laboratory.

PROBLEMS UNDER CONSIDERATION

1. Studies on oxidation in actively growing and differentiated
cells, especially in reference to their growth

In a previous communication (3) before this society evidence
was given by the author to show that connective tissue cells
and simple mesencbyTnatous and embryonic muscle cells show
no evidence of an internal organization suitable for growth.
They act more like homogeneous fluid systems, leaving aside
the nucleus and the astral center. Growth in them can be
explained entirely as a surface tension phenomenon, the energy
of the process being effected by the formation of a substance
within the cell, which is insoluble in circulating body fluid, but
soluble in dead cells, fibrin, and products of disintegrating cells.
A cell forming this substance and brought in contact with fibrin
suffers a decrease in surface tension along the line of contact.
The further details of the organization peculiar to growth is
determined by these contacts. The order of the contacts of
a growing and dividing heart muscle cell is different from a
rhythmically' contracting one.

For a long time it has been known that actively growing
tissue cells, such as those taken from young embryos and
sarcomata commence their growth in the culture within one or
two hours or much earlier than adult tissues. This is seen in
a study of tissues of embryo chicks. The cells from a frag-
ment of the heart of a two, three, or four days old chick embryo
will commence to grow almost at once when placed in the culture.
About fragments of a ten day old embryo heart such activity
is not observed until after six or twelve hours, and about frag-

PROBLEMS IN CANCER RESEARCH

135

ments of a fifteen day old chick embryo heart after fifteen or
twenty hours; while about young adult heart fragments this
latent period is often as long as twenty-four or forty hours.

In a previous publication (4), the writer reported a study of the
relation of oxygen to these changes in the cell. A method was
developed which allowed accurate regulation of the amount of
this gas in the chamber about the cells. The tissues especially
studied were fragments of the heart and body wall of fourteen,
fifteen, and sixteen days old chick embryos. It was found that
the growth was as active in an atmosphere of 9 per cent oxygen
as in pure oxygen, and ceased in atmospheres which contained
slightly less than 6 per cent of this gas. It became of interest
this year to see whether the more actively growing tissue might
not show variations in these figures. The question that arose
was whether the less actively growing tissues did not have to
form or in some other way liberate the substance or substances
which actively change surface tension or produce energy. The
short latent period in the more actively growing cells suggested
a supercharging of this substance or substances.

In 1903, Fletcher (5) showed that a muscle of the adult frog
will give maximum contractions every five minutes for a period
of two hours in an atmosphere of pure nitrogen. In the work on
organization of isolated rhythmical, contracting heart muscle
cells in the tissue culture, the author has found evidence to
show that the chemical changes of muscular contraction are
similar in certain stages to those of growth. If muscle will
contract for a time without oxygen the question that arises is:
Will not cells fully differentiated for growth show the same
reaction (6)?

The tissue so far tested has been cells from the hearts of chick
embryos of various ages, ranging from four to fifteen days.
Fletcher's experiments were repeated with contracting hearts and
heart muscle fragments of these embryos isolated in the tissue
culture. These cells will contract for from twenty to twenty-
two hours in an atmosphere of nitrogen. In the same oxy-
gen-free atmosphere, the cells from fragments of four and five
days old chick embryos will grow from four to six hours. This

136

MONTROSE T. BURROWS

growth commences, however, always after a considerably longer
latent period than that in an atmosphere containing oxygen. It
is active for this short time, after which the cells invariably
disintegrate rapidly. The period of growth in these hanging-
drop cultures is very much longer in air; the early disintegration
is rarely or never seen; and when it occurs, it is always later. To
obtain a similar growth about the fragments of the heart of
ten day a old chick embryos 1.8 per cent oxygen was required and
about fragments of fifteen days old chick embryos 4.5 per cent.

These experiments have not only been applicable to the prob-
lem in question, but have given a method of differentiating func-
tionally the actively growing young embryonic cells from those
showing greater degrees of differentiation. In the case of these
cells they have also shown the necessity of oxygen for the main-
tenance of structure. Whether this method may be used for
differentiating any actively growing tissue is a problem yet to
be considered.

STUDIES OF PLANT CANCERS

Another method which has appealed to us in attacking the
cancer problem has been the study of plant cancers. These
represent an active proliferative growth, the etiology of which is
known. Chambers (7) has undertaken this problem. He was
interested during last summer in investigating the metabolic
activity of the organism, Bacterium tumefaciens, in culture,
and in the course of this study noted that these organisms pro-
duce alkali in the culture even in the presence of sugar, which
they also cause to disappear rapidly. Smith, by a different
method, had noted that ammonia is formed in the cultures of
these organisms. Harvey (8) more recently, also, noted that the
tissue of these tumors is more alkaline than normal plant tissue.
The work of Chambers would indicate, therefore, that the in-
crease in the OH-ion in the tissue of these tumors is due to the
etiological agent, which produces alkali in the cultures at least.
Chambers is now attempting to see if under the condition imposed
on them by the plant, they also form it.

PROBLEMS IN CANCER RESEARCH

137

These facts attracted our attention to the work of Menten
(9), by which an alkalosis was demonstrated in the blood of
cancer patients; and this question is now under investigation.
The various methods of testing were critically reviewed and the
indicator method was selected. The tests are made on a dialy-
sate of venous blood, which is taken under oil and kept unexposed
to air throughout the whole procedure. The results so far ob-
tained have been interesting. In the early part of the disease,
these patients show no evidence of change in the reaction of
their blood when it is taken from a region far removed from the
cancer. A definite alkalosis is observed only when the cancer
has grown to considerable extent and involves neighboring
glands and tissues. The blood of a patient with a beginning
carcinoma in the lip is normal. A definite H-ion change in the
blood is not seen until the glands of the neck are involved, and
then increases gradually as the disease progresses. The same is
true for carcinoma of the intestine. A small annular carcinoma
of the intestine with no or but little glandular involvement may
show by the indicator method no change in the alkalinity of the
blood. Sharp changes are always seen, however, when liver
nodules are present. The C02 relations in the blood of these
patients is now being studied. It is possible that by this means
much concerning the significance of this alkalosis may be deter-
mined. The anemia may also have an effect.

The importance of these experiments is still to be ascertained.
That the alkali is probably produced in the malignant tumor is
indicated, however, in the study of one patient suffering from
a sarcoma of the forearm. Blood taken directly from a vein
leading from the tumor showed a pH of 7.5, while in the other
arm the blood pH was 7.3, or normal. This case illus-
trates the difficulty of using such a method for a clinical test.
The body compensates more readily for acid than alkali, but at
the same time there is a rapid elimination of alkali by the kidney.
In the case of acid, the compensating mechanism gives evidence.
To what extent this can be used for determining the presence of
alkali is yet to be determined. In the case cited, there was
evidently alkali being given into the blood by the tumor, but it

138

MONTROSE T. BURROWS

was not detectable in the other arm. There was no glandular
involvement in this case.

What the further studies of this reaction will open up and the
further studies of the bacteria of the plant tumors may reveal
is a question which the future alone can answer. It seems, how-
ever, that this method may yield something important.

REFERENCES

(1) Stock ard, R. : An experimental analysis of the origin and relationship of

blood corpuscles and the lining cells of vessels. Proc. Nat. Acad, of
Sc., 1915, i, 556.

(2) Burrows, M. T. : Unpublished notes.

(3) Burrows, M. T. : The tissue culture in cancer. Proc. Second Pan American

Sc. Cong., 1915-1916, Sec. viii. Pt. 2, pp. 494-496.

(4) Burrows, M. T. : The oxygen pressure necessary for tissue activity. Am.

J. Physiol., 1917, xliii, 13-21.

(5) Fletcher, W. M. : The influence of oxygen upon the survival respiration

of muscle. J. Physiol., 1917, xxviii, 354-359.

(6) Burrows, M. T. : The reserve energy of actively growing tissues. Proc. Soc.

Exper. Biol, and Med., 1921, xviii, 133.

(7) Chambers, W. H. : Paper before the Am. Physiol. Soc, Chicago, December,

1920.

(8) Harvey, R. B. : Relation of catalase, oxydase and H-ion concentration to

the formation of overgrowths. Am. J. Botany, 1920, vii, 211-221.

(9) Menten, Maud L. : The alkalinity of the blood in malignancy and other

pathological conditions; together with observations on the relation of
the alkalinity of the blood to barometric pressure. J. Cancer Res.,
1917, ii, 179-211.

THE INFLUENCE OF HEREDITY IN DETERMINING
TUMOR METASTASES1

STUDIES IN THE INCIDENCE AND INHERITABILITY OF SPON-
TANEOUS TUMORS IN MICE

Sixteenth Report

MAUD SLYE

From the Cancer Laboratory of the Otho S. A. Sprague Memorial Institute and the

University of Chicago

Received for publication August 1, 192J

Tumor metastases in this stock are somewhat rare. In 29,000
necropsies, furnishing something over 4000 primary spontaneous
tumors, mostly malignant, only about 19 per cent of the tumors
have metastasized. It has been the general opinion of patholo-
gists that the formation of metastases is a distinguishing property
of malignant tumors. The infiltration of tumor cells into
tissues adjacent to the neoplasm, and the entrance of these cells
into the lymph and the blood streams are the usual mechanical
bases for secondary growths. But all these conditions exist in
innumerable instances, both in human and in animal tumors,
without the occurrence of secondary growths. For example,
females 7821 and 9260 of this stock (see chart 11, strain 508,
p. 158) each had a large and highly malignant infiltrative mam-
mary gland carcinoma of relatively long standing. In both
instances there were multiple tumor emboli in the lungs, but no
secondary growths arose from these emboli, — i.e., the mechanical
basis was there in each case, the occurrence of tumor did not
follow.

1 Presented before the American Association for Cancer Research, Cleveland,
Ohio, March 24, 1921.

139

140

MAUD SLYE

Ewing (1), in his chapter on metastasis, speaks of "a peculiar
susceptibility of some tissues to develop secondary tumors, "
but later states that "the mechanism of circulation will doubt-
less explain most of these peculiarities, for as yet there is no
evidence that any one parenchymatous organ is more adapted
than others to the growth of embolic tumor cells/ ' There
would seem, then, to be some factor, other than the mechanical
one, which determines the incidence and location of metastasis,
even though it be minimized, as Ewing has done in the para-
graph cited.

Again, Ewing states that in highly vascular tissues like the
lip, stomach, and testicle, very small carcinoma may yield
distant metastases. But these very organs are among those
whose tumors rarely metastasize in this stock. For example,
among the lip carcinomas in this stock none has metastasized;
among the testicle tumors (2), one metastasized locally, while
five out of six stomach carcinomas (3), (4), showed metastases
in regional lymph-nodes only, but no vascular metastases in the
liver or elsewhere. And this in spite of the close resemblance
of these tumors to human tumors of similar type in similar
organs.

In this stock, mammary gland carcinoma and sarcoma rarely
metastasize into regional lymph-nodes, in marked contrast to
human breast tumors of the same type. Indeed, in this stock,
mammary gland tumors can be seen growing up to the lymph-
node but not invading it, even by extension. On the other
hand, pulmonary metastases from mammary gland carcinoma and
sarcoma are common in this stock, again in contradistinction to
the reported behavior of human breast tumors of the same types.

These divergencies (which are typical of the many divergen-
cies) in the metastasis behavior of tumors of similar types and
in similar organs, require some explanation other than merely
the mechanical tendency of certain types of tumors to metas-
tasize in certain locations, and that there is some other factor is
generally recognized by pathologists, although little progress
has as yet been made in demonstrating what this factor is.

HEREDITY IN DETERMINING TUMOR METASTASES 141

The matter of tumor metastasis in this laboratory has been
under observation for twelve years, and the data here given
are based on over 29,000 necropsies, among which every instance
of metastasis is known. The charts used to demonstrate the
points made in this report are perfectly typical. The non-tumor
fraternities (where there are any) have been omitted, in order to
save space and to show exactly where these tumors occurred.

It must be borne in mind in reading these charts, that, as I
have previously stated, tumor metastases in this stock are
rather rare, being only about 19 per cent of the primary malig-
nant tumors. A given chart then, which shows many primary
tumors, must not be expected to show large numbers of secondary
tumors. For example, a chart showing 8 primary tumors,
would, if this general average of 19 per cent were carried out,
show only 1.5 cases of metastasis, etc. The biological evidence
from this laboratory, which explains this low percentage of
metastasis from malignant tumors, will be given further on in
this report.

From this study three points stand out with striking clearness:

POINT I

In any given strain the metastatic tumors (where there are any)
tend to occur in exactly the same organs in which the primary tumors
of that strain occur.

Note chart 1, showing part of strain 215 and some of its deriva-
tives. The parent female, no. 3, had a sarcoma-carcinoma of
the mammary gland, a malignant adenoma of the liver, and a
metastatic sarcoma of the kidney. The parent male, no. 360,
came from a strain carrying lung and mediastinal tumors, and
he himself was proved heterozygous to tumors in these locations
(i.e., he carried them potentially and introduced them in every
strain into which he was crossed, although he did not himself
show these tumors).

Note here how these particular types and locations of tumors
have " segregated out" and certain of them have been trans-
mitted to the branches of the strain here shown, both as primary

HEREDITY IN DETERMINING TUMOR METASTASES 143

and as secondary neoplasms. In the first hybrid generation,
male 1192 had a primary carcinoma of the lung and a secondary
carcinoma of the lung. Female 5 had a primary carcinoma of
the mammary gland, with multiple secondaries in the lung.
Note how the different types and locations of tumors have
"segregated out/' so that in branch A lung tumors alone have
been transmitted, and a 100 per cent lung tumor strain is being
extracted. In branch B liver tumors have predominantly
been transmitted, from female 630 with a secondary tumor in
the liver, so that in one part of the branch a 50 per cent fiver
tumor strain is being extracted.

In these two branches of the strain, then, female 3 introduced :

a. Primary carcinoma of the mammary gland, in females 5,
2239, and 630.

b. Primary adenoma of the liver, in females 1681, 8871,
11237, and 12207, and in male 2327. (Note the three consecu-
tive generations of adenoma of the liver.)

c. Secondary carcinoma of the liver, in female 630.

d. One case of sarcoma, in female 5543.
Parent male 360 introduced:

a. Primary carcinoma and adenoma of the lung, in males
1192, and 900, and in females 326, 816, 823, 3191, 2548, and 6688.

6. Secondary carcinoma of the lung, in male 1192 and in
females 5 and 630.

c. Primary tumor of the mediastinum, in female 5543.

d. Secondary tumor of the mediastinum, in female 326.

Note how female 5543, with a lymphosarcoma of the medias-
tinum, derived the type of her tumor from her maternal ancestor
(female 3) and the location of her tumor from the paternal
ancestor, male 360 (who carried these tumors potentially, and
transmitted them in every strain into which he was crossed,
although he himself did not develop them), just as in the human
species, a child might inherit curly red hair from a mother with
curly black hair, and a father with straight red hair.

Note how female 5 derived her primary tumor from the
maternal ancestor, and her secondary tumor from the paternal
ancestor, and how female 630 derived her primary tumor from

144

MAUD SLYE

the parent female, the secondary in the lungs from the parent
male, and the secondary tumor of the liver from the parent
female — a marked demonstration of the " segregating out" of
these things, which is the heredity test. Note how, after the
second hybrid generation, no further mammary gland tumors
occurred in either of these two branches.

Note that there occur no tumors, either primary or secondary,
which were not introduced by the parents. Note that the
secondary tumors occur in the same organs as the primary
tumors.

Chart 2 shows parts of strains 48 and 292. These are two
strains derived from the same original mating as was strain
215, that is, female 3 (with a sarcoma-carcinoma of the mam-
mary gland, a malignant adenoma of the liver, and a metastatic
sarcoma of the kidney), and male 360 who introduced carci-
noma of the lung and the mediastinum. Here again the particu-
lar types and locations of tumors introduced by the parents
have " segregated out" and been transmitted to the resulting
strains, both as primary and as secondary neoplasms.

In the first hybrid generation female 5 derived from the
mother a carcinoma of the mammary gland, and from the
father a secondary carcinoma of the lung. In the second hybrid
generation, female 37 derived from the parent male a carcinoma
of the lung, with a secondary carcinoma in the mediastinum.

In the third and fourth hybrid generations, note the out-
cropping of sarcoma, both primary and secondary, in females
1454, 26, 348 and 396. Liver tumors derived from female 3,
also occur in these strains, as shown in females 5743, and 26, and
in male 11836. There is also a " pre-cancerous " liver in female
399. Here also lung tumors, derived from male 360, occur
primary in females 37 and 399 and in male 11836; and secondary
in females 5, 5743, and 396.

Note the occurrence of secondary carcinoma of the mediasti-
num in female 37, of secondary sarcoma of the mediastinum in
female 396; of secondary carcinoma of the lung in female 5, and
of secondary sarcoma of the lung in females 5743 and 396. Note
the three cases of secondary sarcoma in the spleen in females

HEREDITY IN DETERMINING TUMOR METASTASES

9*9,

9 < S

o Is:

<J 2 J

31,

«0

«o

2 o

I | J»

< < -

c «

2<o
■

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 2

146

MAUD SLYE

5743, 348, and 396. Rarely in this stock, except in strains
derived from female 3, have there been any tumors of the spleen
either primary or secondary.

Here again, after the first hybrid generation, mammary gland
tumors disappear — another indication of how these things
" segregate out." The mammary gland tumors preponderated
in other strains derived from the same original mating. The
dominating tumors, both primary and secondary, in these
strains are in the liver, kidney, spleen, mesentery, lungs, and
mediastinum.

In these strains, then, female 3 introduced:

a. Primary carcinoma of the mammary gland, in female 5.

b. Primary adenoma of the liver, in female 399 and in male
11836.

c. Primary sarcoma of the liver, in females 5743 and 26.

d. Primary sarcoma of the kidneys, in females 26, 348, and 396.

e. Secondary sarcoma of the kidneys, in female 1454.

/. Secondary sarcoma of the spleen, in females 5743, 348, and
396.

g. Primary sarcoma of the mesentery, in female 1454.

h. Secondary sarcoma of the mesentery, in females 1454,
5743, and 26.

Parent male 360 introduced :

a. Primary lung tumors, malignant and not yet malignant,
in females 37 and 399, and in male 11836.

b. Secondary lung tumors, in females 5743 and 396.

c. Secondary carcinoma of the mediastinum, in female 37.

d. Secondary sarcoma of the mediastinum, in female 396.
Note here again, how, though they are much fewer, the second-
ary tumors arise in the same organs as do the primary.

Chart 3 shows a part of strain 65, branch 2. The parent
female, 5738, had three carcinomas of the mammary gland
and metastatic tumors in the lungs. She was mated with her
brother, male 782, who died of uncertain causes, but who was
proved heterozygous to tumors of the mammary gland (he
carried them potentially and transmitted them, but did not
himself develop them). Note here that with one exception

HEREDITY IN DETERMINING TUMOR METASTASES

J o

o-1
£|

y liO

Of ui IT)
< Q OJ
U< ~

< "> cO

_l2

Chron.

ir>

j
O

148

MAUD SLYE

(female 9888 who had a lymphosarcoma of the thymus) the
only tumors which occurred were carcinoma of the mammary
gland and lung tumors, primary and secondary.

Note the five consecutive generations of carcinoma of the
mammary gland in branch B, all but females 5753 and 9903
showing lung metastases. Note the heavy percentage of lung
tumors primary and secondary in both branches of the strain.
Note the negative evidence of the " segregating out" of the
different types and locations of tumors, both primary and
secondary, in the complete absence from this strain of all tumors
of the liver, kidney, mesentery, spleen, pancreas, uterus, etc.

Chart 4 shows a part of strain 338, branch V, A, with partial
ancestry. The parent female 3920 had an adenoma of the liver.
Note the predominance of primary liver adenoma in this strain,
7 cases (over 43 per cent) and the occurrence of two cases of
secondary liver tumor in the same strain, that is, female 8865
with an osteoid sarcoma of the mammary gland, secondary in the
liver, and her grandson, male 16370 with an osteosarcoma of the
leg, metastasizing in the liver, and also a mesothelioma of the
testicle. (Testicle tumors also occur in this strain.)

Chart 5 shows a continuation of the same strain, strain 338,
branch V, A (the data being too extensive for a single chart).
Chart 5 begins with generation 3 of the family, the mating being
between female 8619 and male 8751. Generations 7, 8, and 9
are added to the family in this chart. Note here, also, the
further occurrence of liver tumors, there being one additional
primary, in female 30469, who showed a primary liver adenoma
in addition to seven carcinomas of the mammary gland, a pri-
mary lung adenoma, a typical uterine fibroid, and a diffuse
endothelial hyperplasia of the glands and spleen. (Needless to
say, the mouse was riddled with tumor, little but neoplasm
remained.) There were also in this part of the family, two
additional cases of secondary liver tumor, one carcinoma and one
sarcoma, that is, female 22263 with three carcinomas of the
mammary gland and a sarcoma in the peritoneum near the
stomach, with sarcoma metastases in the liver, pancreas, and
uterine wall, and female 30501 of generation 9, with two carci-

HEREDITY IN DETERMINING TUMOR METASTASES 149

6s i
*o

150

MAUD SLYE

nomas of the mammary gland and the liver absolutely riddled
with carcinoma secondaries.

It is interesting to note here also the two occurrences of uterine
neoplasm, primary in female 30469, and secondary in female
22263, as this is one of the strains carrying most of the uterine
tumors in this stock.

CONTINUATION OF STRAIN 338 — BR-2T A«^»

G t

2 Cahc (V) C

Adenoma Liver

2>j «75 1

1 2 Caixc (vyGi.

I Adenoma LivEf

1 9112

ri 102.50

Carc f*) Gi.
lOOfefe

Adenoma Livtn
lll$0

2 Carc.M. Gu
5aHC M Ci
12201

CaR.C Li.
.2 12 <J1

Iadcn-o^a Long

P Leukemia

JCarx. M Gi».— PediTONtAL Sar.c.

Meta* L.VER. - PamCR.EAS

171.81

Sj 19913

222 UTER.INE VVal.l

Ca*c ft* Cl

laTZ7H

SaV

o^aa Head

£ <. 2.5-9

7 Carom. Gl — liver. ade*oa\a

Luwc Adinomh — U-TEMne Fibroid
DlFFuSE EmDOtm. Hypen.PL Mia of
30 H (.9 Gl-AMOSi and S PL 1 £n

Chart 5

2. C arc . M. Gi..

l-iV«B. N£»HLt f.ti'ncio

fir AABTAS

The pancreatic secondary sarcoma also is extremely interest-
ing, as this is the strain which carries the only other pancreatic
neoplasms which have occurred in this stock.

The complete ancestry of this strain is shown in charts 6 and
7, chart 6 being the ancestry of parent female 3920, and chart 7
the ancestry of parent male 3024. The prominence of liver
tumors, primary and secondary, in the resulting strain is due to
the fact that it was bred in from both sides of the family; the

HEREDITY IN DETERMINING TUMOR METASTASES 151

ANCESTRY OF Q 3920 — STRAIN 338

Carc. Sarc.M Gu
AWig. Adenoma Liver ^
(^13 Sarc.Metas KiDNiev (J

Cmron. Nephfiitis
30

Carc MGl Sup. Panopthalmitis

1 Caro.AA.Gu

Intestinal Infection

Meta^. Lungs

O "791

1481

|Adenoma Liver.
£|3920

Chart 6

strain thus received "a double dose. " Note also the marked
outcropping of sarcoma in strain 338, derived from original
parent female 3. Sarcoma also was bred in from both sides of
the ancestry.

ANCESTRY OF 6 3Q2A - STRAIN 358

I Carc. Sarc.M.Gl.
Malic-. Aoenon\a Liver
Q I 3 5arc. AAetas. Kioncv

I X

IAdenonva Liver.
_ 883

I Chron Nsphritis
6 30

Uncertain

6 I 842
I

Chron. Nephritis Ur'NAIVy Reten
A Wounds

6 1 ion

9 1441

Chron. Nephritis

Inflam.Lung Nodule

3024

Chart 7

Chart 8 shows a part of strain 465, which was derived from
the same original ancestry as strain 338 (ancestry shown in
charts 6 and 7). Note here also, the same appearance of liver
tumors, two primary and one secondary. This gives us, then,

152

MAUD SLYE

in these two small strains (338, branch V, A and 465) derived
from identical original ancestry, 10 primary and 5 secondary
liver tumors. Including the ancestry, the number is 13 primary
and 5 secondary liver tumors. This is a very high occurrence
of liver tumors, both primary and secondary, especially when one
considers the fact that outside of this stock there is just one
mouse liver tumor on record, a primary tumor of the liver, re-

PART OF STRAIN 465

Adenoma Liver |Chron (Nephritis

3920 6 13024-
« I

1 Ch«on. Neph.

1 Uncertain

Q 15333

612512

Sarcoma Liver.

5305

IChron. Neph.

6 15215
I

Accident

1 Uncertain

5 15885

6 I&369

Sa«c. Hind Leo Uncertain I Chron. Ncphrit.s

R-etrop. Sarc- Metas a

almost replaonc Q'9957 Ol 88£8

9 979 E..VER T I I

Sarc. Kidme*

Chron. Nephritis

Abscess Lung

9847 6

8852

Lymphoio Tumor Sto/mach
Lympho- Sarcoma L.ver
11252 anoCommjn &iui

Chart 8

ported by the Imperial Cancer Research Fund of England. This
liver tumor occurrence (in the Slye stock) within a single strain,
makes a piece of evidence which is conclusive.

These five strains, represented in charts 1 to 8 inclusive (which,
it must be remembered, are perfectly typical) furnish both
positive and negative evidence of the " segregating out" of differ-
ent types and locations of tumors, and their transmission as
such, whether as primary or as secondary neoplasms. They

HEREDITY IN DETERMINING TUMOR METASTASES 153

furnish, therefore, both positive and negative evidence that
in any given strain the metastatic tumors (where there are any) tend
to occur in exactly the same organs in which the primary tumors of
that strain occur. Note, therefore, that :

1. Every secondary tumor of the liver in this stock (with the
possible exception of two osteosarcomas, which metastasized into
nearly every organ in the body, in two mice whose strains have
not yet been completely analyzed) has occurred in strains carry-
ing primary liver tumors.

2. Every secondary tumor of the kidney in this stock has
occurred in strains carrying primary kidney tumors.

3. Every secondary lung tumor in this stock has occurred in
strains carrying primary lung tumors (with the possible exception
of the one lung metastasis in strain 164, branch IV, line C).

4. Every secondary tumor of the mediastinum in this stock
has occurred in strains yielding primary tumors of the medias-
tinum.

5. Every secondary tumor of the pancreas in this stock has
occurred in the strain which yields primary pancreatic tumors.

6. Every secondary tumor of the spleen in this stock, has
occurred in strains carrying primary spleen tumors.

7. Every secondary uterine tumor in this stock has occurred
in strains carrying primary uterine tumors.

When this stock, or any other, has been completely analyzed,
there is little doubt that it will show that every secondary tumor
of any organ whatever has arisen in a member of a strain which
has shown, or will show, some cases of primary tumors in that
organ (however few they may be). There is little doubt that
strain 164, branch IV, line C, of this stock will in time yield a
primary lung tumor, although it has not done so yet. There is
a very slight inheritance of lung tumor neoplasm tendency in
this line of the strain, and it is therefore to be expected that
lung tumors of any kind will occur but seldom.

The complete analysis of any strain is a very long and difficult
process, and the apparent testimony of the frequent occurrence
of secondaries in man in tissues where primary tumors rarely
occur is of no help here, as no human strain has ever been even
partially analyzed.

154

MAUD SLYE

Female 3 is listed in these charts as having a sarcoma-carci-
noma of the mammary gland. This tumor was unquestionable
carcinoma in some parts and apparently sarcoma in other parts.
There has been considerable difference of opinion concerning
these mixed tumors. Slye, Holmes, and Wells (5) Loeb (6),
Lewin (7), Bashford (8), Woglom (9), and some others, diagnose
them as such, while Le Count (10) and Ewing (11), etc., consider
them to be entirely carcinoma, with pressure distortion of the
cells simulating sarcoma in some areas.

The biological evidence of the work in this laboratory un-
questionably supports the opinion that there are these mixed
tumors. Female 3 with such a tumor, has unquestionably
transmitted carcinoma in every strain into which she has been
hybridized; and she has equally certainly transmitted sarcoma
in some branch of every strain into which she has been crossed
(see charts 1, 2, 4, 5 [female 3 shown in ancestry charts 6 and 7],
8 (same parentage charts 6 and 7) , Chart 9) .

I have preferred the term sarcoma-carcinoma to represent
this type of tumor, because according to the biological evidence,
it is not a sarcomatous carcinoma, nor a carcinomatous sar-
coma, but rather a sarcoma plus a carcinoma, each type of tumor
being transmitted separately, as such. Biological evidence, as
manifested in heredity behavior, is too fundamental to be ignored.

POINT II

In certain strains there is a tendency for primary tumors to
metastasize in certain organs; whereas, in other strains, tumors of
the same type, primary in the same organ, even when they are of
older and of larger growth, fail to metastasize into those organs.

Chart 9 shows part of strain 73, which was derived from the
same female 3, mated this time with male 30. Male 30 came
from a strain carrying tumors of the lung, mediastinum, and
diaphragm, and was proved heterozygous to tumors of these
organs, having been tested in various crosses.

The son of this mating, Jap. W. F. male, died before autopsies
were made, and consequently the cause of his death is not known.

HEREDITY IN DETERMINING TUMOR METASTASES

O < ? x —

J) go
Of

> Q

>. O cC

5* »

U g <\

? 2

u >

-1 Ul

S; 2

_1 a

2 *
ul

< a

< ^

x £ in

* i

Ul J K_

-1 O, E

£ g 10

156

MAUD SLYE

He was crossed with a Lathrop silver-faun female, who also
died before autopsies were made. Their offspring, however, are
shown in two branches, A and B.

Note how in branch B, every mammary gland carcinoma
metastasized into the lungs; even the retroperitoneal and pelvic
sarcomas (male 3951) metastasized in the lungs. Note also
that with the exception of male 1775 this branch shows 100 per
cent of lung tumors, primary and secondary. Note the two
generations of metastasis into the diaphragm, female 1722 and
male 3098.

On the other hand, in branch A, carcinomas of the same type
and of older and larger growth, failed to metastasize into the
lungs (females 4328, 5723, and 7544). Note that in female
5723, although the mechanical basis for lung secondaries was
present in multiple tumor emboli throughout the lungs, no
lung tumor developed. Note that the sarcoma of the mesen-
tery in male 4630 metastasized into the liver, kidney, and retro-
peritoneal glands, but not into the lungs.

Note the three consecutive generations of leukemia in this
strain, in males 899, 1785, and 2360; and the further fact that,
although the lungs are one of the principal organs for leukemic
infiltration, in males 1785 and 2360 there was no infiltration in
the lungs, while in male 899 the lung infiltration was very slight.
On the other hand, the marked leukemic infiltration in this
family was in the same organs in which the tumors of the parent
female occurred, that is, the liver and kidney. The metastatic
tumors, also, in this family were in these same organs.

In branch B, however, the 100 per cent lung tumor strain,
female 6568 who had pseudoleukemia along with an adenoma
of the lung, showed marked infiltration of pseudoleukemia
throughout the lungs.

We have here, in these two branches of the same strain, a
striking illustration of point II: In branch B all tumors metas-
tasize into the lungs, and pseudoleukemia picks out the lungs
predominantly; while in branch A, the tumors of the same type
in the same organ and of older and larger growth fail, in every
case, to metastasize into the lungs. Even the leukemic infiltra-
tion fails to take hold of the lungs in this branch.

HEREDITY IN DETERMINING TUMOR METASTASES

157

Chart 10 shows part of strain 405. In this branch of the
family, after the parent generation, although there are five
straight generations of mammary gland carcinoma, none of these
tumors metastasized into the lungs. Note also the complete
absence of primary lung tumors. This branch of the family

PART OF 5TRA1N 405

6.i

3 Carc. M.Gl.

Uncertain

AAetas. Lungs

5738

782

^PARENT

Carx.M.Gl.

Pul. Infection

Ch rcn. Mast it fs

5753

2715

Garc. M.Gl.
5581

Carc.M.Gl.

Chron. Neph,

Pe-ritonitis

4449

Carc.M.Gu.
Tapeworm .
G0&2 O

Wounds
4,571

Wounds
10342

Or. .

Carc.M.Gl.
12G&D

Chart 10

did not inherit lung tumors, and none occurred, either primary or
secondary.

Chart 11 shows one branch of strain 508. Here again through
five consecutive generations of mammary gland carcinoma, no

158

MAUD SLYE

lung metastasis occurred. Even where the presence of multiple
tumor emboli in the lungs gives certain evidence that cells from
the primary tumor were carried to the lungs (in females 7821 and
9260) no lung tumor occurs. The mechanical basis was there;
the tumor growth did not follow. This strain also shows no
primary lung tumors.

PART OF STRAIN -508

•PARENT

2 Carc.M.Gl.
59b5

Occlusion or stomach
t9G8

'parent

CARC.rA.GL.

Tumor Emboli Lungs
7821 O

Chron. Nephritis
5212

Carc.MGl.
7237

Wounds
701 T

Carc.MGl.

Tumor Emboli Lungs

Accident
74-80

?Carc./A.Gl.

9989

Chart 11

Chart 12 gives a summary of strain 164, branch IV, families
A, B, and C. It is shown in summary only in order to get the
data within a single chart. These branches of the family showed
48 primary tumors, all but one of them highly malignant, with
only one case of lung metastasis, in female 10574, a member of
line C. Note the marked absence of lung tumors primary and
secondary, although lung tumors are second in frequency of all
mouse tumors.

HEREDITY IN DETERMINING TUMOR METASTASES

The strain also showed 13 cases of pseudoleukemia and one of
leukemia. Here again, it is most interesting to note that the
six cases of pseudoleukemia which occurred in line B showed not
the slightest infiltration in the lungs. The malignant thymus

STRAIN 164- -BR.TST-A-B-C

PRODUCED THE. FOLLOWING CASES OF TUMOR

WITHOUT METASTASIS

WITH METASTASIS

Carcinoma M.Gl.

1 IN LUNG

Sarcoma W.Gl.

Ovarian Adenoma

Malig. Thymus Lymphoma
Did not Infiltrate lun&s

LfMPHO - v5ARC0MA Thymus

I>I3> NOT INFILTRATE LUNC-5

5q.CEi.L Cab.c. Back

v5r Cell vSarc. Back

S(^.Cel.l- Carc Face

5p C stl i_ Sarc.Face

Retropfr.it oneal 5arcoma

Total

P5EUDO Leukea^ia

Note the absence or primary lung tumor and
THE Single Case of lung metastasis

Chart 12

tumors in this line of the family also failed to infiltrate the lungs,
although they went extensively into the chest wall and medias-
tinal glands. The seven cases of pseudoleukemia which fell in
line C of this strain, showed a very slight infiltration of the lung.
This is the same line which showed one case of lung metastasis.

160

MAUD SLYE

We have then, in this strain, lines A and B, a marked illus-
tration of the failure of the lungs to yield at all to tumor, primary
or secondary. In other words, the lungs of this family do not
show neoplastic proliferation, nor do they accept neoplastic
secondaries or leukemic growth in spite of the certain and marked
presence of the ordinary provocation for such proliferation. In
line C of the strain, a slight tendency of the lung to proliferation
is present.

This strain was derived by the hybridization of an absolutely
nontumorous strain of house mice, in my hands many years,
strain 358, with albino strain 146, which carried many tumors,
and furnished some 100 per cent tumor families. This strain
gives another marked illustration of the segregating out of these
things; the tendency to neoplastic lung proliferation, either
primary or secondary, failed to be transmitted at all to lines
A and B of branch IV of the strain, and only slightly to line C.
Note the striking contrast between this strain with no lung
tumors, and strain 522 (chart 15) showing 100 per cent lung
tumors (p. 167).

These four strains represented in charts 9 to 12 inclusive
(which are perfectly typical) furnish both positive and negative
evidence of the segregating out of different types and locations of
tumors, and their transmission as such, whether as primary or as
secondary neoplasms.

They furnish therefore evidence that in certain strains there is
a tendency for tumors to metastasize into certain organs, whereas in
other strains, tumors of the same type in the same organ, even when
they are of older and of larger growth, fail to metastasize into those
organs. Moreover in these strains, leukemia and pseudoleuke-
mia tend to infiltrate the same organs which show primary and
secondary neoplasms, and fail to take hold of the same organs
which do not yield to primary or secondary tumors.

In this connection, it is interesting to analyze the results of
various experiments to test the effect upon metastasis formation
of making exploratory incisions into tumors, of massaging
tumors, etc. Let me indicate what would be a complete biolog-
ical control in such experiments.

HEREDITY IN DETERMINING TUMOR METASTASES 161
TWO ILLUSTRATIONS OF COMPLETE BIOLOGICAL CONTROL.

Strain 164 branch C of the Slye stock, during its twelve years
of existence has consistently yielded 0.01 per cent of lung tumors
primary and secondary, leukemic and pseudoleukemic infiltration
of the lung, and infiltration into the lungs from malignant thymus
tumors.

If now we use tumorous mice from strain 164 branch C, and
excise for diagnosis pieces of their tumors or massage their tumors
or otherwise manipulate them, and if we find at autopsy that the
group of mice whose tumors have been manipulated show little
or no higher percentage than 0.01 per cent of lung tumor metas-
tasis, and that the group which is not manipulated runs true to
this 0.01 per cent lung tumor metastasis, we shall have a perfect
control, and shall be justified in saying, without qualification,
that excising a piece of tumor for diagnosis or massaging the
tumor, or otherwise manipulating it, had no effect upon the pro-
duction of lung tumor metastasis.

Strain 65 branch II B has during its twelve years of existence
produced 35 per cent of lung tumor metastasis.

If now we repeat these same experiments with this strain, and
find at autopsy that the group whose tumors have had a piece
excised for diagnosis or which have been in any other way manip-
ulated, show little or no increase in lung tumor metastasis over
35 per cent, and if the unmanipulated group runs true to this
percentage of lung tumor metastasis, we shall have a perfect
biological control of the experiments, and shall be justified in
saying without qualification that the excision of a piece of tumor
for diagnosis or any other form of manipulation had no effect
upon the production of lung tumor metastasis.

162

MAUD SLYE

THREE ILLUSTRATIONS OF THESE EXPERIMENTS WITH INADEQUATE
BIOLOGICAL CONTROL

Strain 164 branches A and E have during their existence
for twelve years, without exception of any sort, yielded 0 per
cent of lung tumors primary or secondary even where tumor
emboli were present in the lungs, 0 per cent of leukemic or pseudo-
leukemic infiltration into the lungs, and 0 per cent of infiltration
into the lungs by extension from thymus, diaphragmatic, or other
adjacent malignant tumors.

If now we repeat these same experiments with these branches
of the strain, we shall find at autopsy 0 per cent of lung tumor
metastasis in every group of mice, whether or not we excise a
piece of tumor, or massage the tumor, or otherwise manipulate it.
That is, a strain whose lungs have been proved not to yield to
tumor emboli is not a strain fitted to show the effect of these or
any other types of manipulation.

Strains 139, 522, 73, branch B, etc., have during their many years
of exi-tence without exception furnished 100 per cent of lung
tumors, primary and secondary, infiltration into the lungs by
leukemia and pseudoleukemia, and extension infiltration into the
lungs from malignant thymus and other adjacent tumors.

If now we repeat these same experiments with these or any
other 100 per cent lung tumor strains, we shall find at autopsy
an exceedingly high percentage of lung tumor metastasis, prob-
ably 100 per cent in every group whether excised or not. That
is, a strain which for twelve years has shown 100 per cent of lung
tumor is not a strain fitted to show the effect upon lung metastasis
of this or any other type of manipulation.

Ill

Groups of mice even when purchased from one dealer are mixed
lots, secured from many divergent sources, and wholly unanalyzed

HEREDITY IN DETERMINING TUMOR METASTASES 163

as to tumor potentiality. They may include some individuals
from 0 per cent lung tumor strains, some from 100 per cent, some
from 35 per cent, etc. If we take such groups of mice and repeat
these same experiments, we shall have no biological control, and
we shall not be justified in saying whether the lung tumor meta-
stases found at autopsy were or were not caused by the excision
of a piece of tumor for diagnosis, or by massage, or whatever
form of artificial manipulation may have been used.

It is greatly to be deplored that the vast majority of experi-
mental studies in animal pathology and bacteriology have not
been and are not as yet being conducted with animals whose
hereditary potentialities have been studied. It is being taken
for granted that pathological conditions found in the' animal
after experimental procedure, are the result of such experi-
mental procedure. No allowance is being made for heredity.

Until stocks of animals to be used in such experiments, have
been thoroughly tested out as to their inherited potentialities,
such experiments are lacking in adequate control, as the factor
of heredity is not being considered, although it is tremendously
potent. Heredity alone would be sufficient to give the exact
results attained in countless experiments which have been con-
ducted without any effort to study or control the heredity
factor. Moreover, most of these studies in metatasis production
by the artificial manipulation of the primary tumor have been
conducted with grafted tumors. As has frequently been stated
in the series of cancer studies from this laboratory, the biological
difference between spontaneous and grafted tumors is so funda-
mental and so complete, that the behavior of grafted tumors has
practically no bearing upon the behavior of spontaneous tumors.

POINT III

Individuals with secondary tumors in any given organ seem to be
as potent as individuals with primary tumors in the same organ to
transmit by heredity primary tumors in that organ.

Note chart 13, showing part of strain 304 with ancestry. Here
two successive generations of females (529 and 467) with second-

164

MAUD SLYE

HEREDITY IN DETERMINING TUMOR METASTASES 165

ary lung tumors, have originated a strain which, in line C, is
producing a 44.4 per cent primary lung tumor strain. The
combination of both lines B and C shows 33.3 per cent of primary
lung tumors. Note the 100 per cent pseudoleukemia family
being extracted in branch B, in all of which the lungs were infil-
trated extensively.

Again, in chart 14, strain 338 branch VI, a 100 per cent lung
tumor strain is being extracted from the original mating of female

PA RT OF STRAIN 3 36 —

2Ca«.m,Gl. I AHMMA Lun e

ARTfRIAL StLtROSlS

2 Cahc A*. Gl.

jCAHC HGl.ONTiror

Sc^. bu Cahc. Jaw

5^.CeuL Carc.

I07«»3

Zarc. Lung Ca*.c.A\Gl.

J Aiu4.T. Lung MtTAS.

5| 9763 e«.m

Carc M.Gl. 1 3 Cakc. M.Gl.

j Adino^aa

! T C arc M Gi..

Lungs Riddled with | aab-av h/n«s

i LUN(

| Mult LumC «fTA»,

12050. AA6TAS ^ j 15161 ^

2 Cakc a* Gtk.
9115190 UTe<i»t r,8w«,

Chart 14

5417, with a secondary lung tumor, with male 7736, with a
primary lung tumor; that is, the metastatic lung tumor is behav-
ing like a primary lung tumor in helping to produce a 100 per
cent lung tumor strain, primary and secondary.

Refer again to chart 1 (page 142), showing part of strain 215,
and note how female 630 (parent of branch B) with a metas-
tatic carcinoma of the liver, produced three straight generations
of primary liver adenoma (not yet malignant) females 8871, 11237,

166

MAUD SLYE

and 12207, as well as a primary malignant adenoma of the liver
in male 2327. Note also how her secondary lung tumor was
followed by a primary lung tumor in female 823 (G3) .

Refer again to chart 2 (page 145), showing strains 48 and 292,
and note how the secondary sarcoma of the kidney in parent
female 3, resulted in primary sarcomas of the kidneys in females
26, 348, and 396. Refer again also to chart 3 (page 147), strain
65, and note how secondary lung tumor in the parent female 5738,
introduced primary lung tumors in the strain, in females 1253,
8749, 11512, and 13386, and in male 11832. Note that the thy-
mus sarcoma in female 9888 was markedly infiltrating the lungs,
while in strain 164 Br. IV, lines A and B, which did not produce
lung tumors of any sort, either primary or secondary and in
which leukemia and pseudoleukemia failed to infiltrate the lungs,
the thymus tumors also failed to infiltrate the lungs.

Note also how in chart 15, strain 522, female 5636, with a
secondary lung tumor (G2), mated with male 8102 with a primary
lung tumor, produced a 100 per cent primary lung tumor strain
(center of the chart). The pseudoleukemia in this 100 per cent
lung tumor strain, viz., female 14033, showed marked infiltration
in the lungs.

From these charts, viz., 13 to 15 inclusive, and also from charts
1, 2, and 3 (which are perfectly typical) it is evident that individuals
with secondary tumors in any given organ seem to be as potent as
individuals with primary tumors in that organ, to transmit by
heredity, primary tumors in that organ.

DISCUSSION

We have here, in the charts presented in this report, biological
evidence of the nature of cancer heredity. It is well to bear in
mind that until the facts of pathology have been based upon the
more fundamental biological facts, it will never be possible to get
a complete explanation of the pathological behavior of tissues.

Let me repeat at this point, what I already have stated fre-
quently in previous publications: the materials used for this
study are individuals analyzed as to their hereditary poten-

HEREDITY IN DETERMINING TUMOR METASTASES 167

2 «f

0
2

o
z

_j !

I I

3
— 1

Sit

I I ol
Q D °

< >• o

< < (O

U f

6 I

£ i

^ 00

a < J>

< O

c3 5

o
z

j-
o

•o

£C_2
Of

168

MAUD SLYE

tialities by every possible test. The tumors are all spontaneous,
arising without interference of any sort except that of selective
breeding. By the process of selective breeding alone, using
analyzed individuals as material, it is possible to extract strains
in which the lungs do not react with either primary or secondary
neoplastic proliferation or leukemic infiltration; or strains in
which the liver does not so react, or the kidneys, or the mammary
gland tissue, etc.

Etiological meaning. The fact that both primary and second-
ary tumors of a given organ or organs tend to occur in the same
strains, and fail to occur at all in other tumor strains, indicates
beyond question that heredity is a strong factor in determining not
only where the primary tumors shall occur, but also where the
secondary tumors shall occur.

The consequent fact, that it is possible to extract strains
where only certain organs shall furnish both the primary and the
secondary neoplasms, and shall yield to leukemia and pseudo-
leukemia; and to extract other strains in which these organs
never furnish either primary or secondary neoplasms nor yield
to leukemic or pseudoleukemic invasion, and that secondary
tumors are just as potent as primary tumors in the transmission
of primary tumors in any given organ, shows beyond a doubt
that the thing which is transmitted in the heredity of cancer is the
tendency of a given organ or organs to yield to cancer. This
inherited tendency of an organ or organs to yield to cancer is
manifested, whether the lesion is primary in that organ, or
whether cells from the primary growth lodge in that organ and
form a secondary lesion.

The negative evidence on this point is just as conclusive,
that is, that in strains from which tumors of a certain organ or
organs have been eliminated by heredity, the cells from the
primary neoplasm fail to take hold, even when they lodge in
such organ, and do not form a secondary growth; and even when
a tumor completely surrounds an organ, it fails to penetrate it
by extension.

Now what do these things mean? They mean that the ten-
dency to carcinoma segregates out and is transmitted as such;

HEREDITY IN DETERMINING TUMOR METASTASES 169

that the tendency to sarcoma or adenoma, etc., segregates out
and is transmitted as such; that a strong tendency to the location
of one or more of these types of tumor in a specific organ, such
as the liver, the lung, the kidneys, the mammary glands, etc.,
is transmitted, owing to the segregating out of a peculiar type of
tissue in these organs, which will respond in a neoplastic manner
to lesions of any kind which furnish a chronic irritation of not too
destructive a type. That is, the tissues of these organs are of the
same nature as the ancestral organs from which they are derived.
There is a specificity of tissue type from liver to liver, or from kidney
to kidney, etc., which will make those organs react in a given way to
a given type of irritation.

Now with each of these things being transmitted as such, it is
possible for a parent female with a sarcoma-carcinoma of the
mammary gland, a malignant adenoma of the liver, and a second-
ary sarcoma of the kidney (like female 3) to transmit to the
strains derived from her, a tendency to:

1. Carcinoma of the mammary gland.

2. Sarcoma of the mammary gland.

3. Adenoma of the mammary gland.

4. Carcinoma of the liver.

5. Sarcoma of the liver.

6. Adenoma of the liver.

7. Carcinoma of the kidney.

8. Sarcoma of the kidney.

9. Adenoma of the kidney, or any combination of these nine.
If she is mated with a male, either himself having a lung and a

mediastinal tumor, or being heterozygous to these locations of
tumor (as in the case of male 360) , the resulting strains will show :

1. Carcinoma of the mammary gland.

2. Sarcoma of the mammary gland.

3. Adenoma of the mammary gland.

4. Carcinoma of the liver.

5. Sarcoma of the liver.

6. Adenoma of the liver.

7. Carcinoma of the kidney.

8. Sarcoma of the kidney.

170

MAUD SLYE

9. Adenoma of the kidney.

10. Carcinoma of the lung.

11. Sarcoma of the lung.

12. Adenoma of the lung.

13. Carcinoma of the mediastinum.

14. Sarcoma of the mediastinum.

15. Adenoma of the mediastinum, or any combination of these
tumors.

The neoplastic growth may be primary in any one or more of
these organs, according to where there is occasion for the primary
lesion to occur, and there may be secondary growths in any one or
more of these organs, according to where the secondary lesions
occur.

Now note that this is exactly what we get from the matings of
female 3 (with a sarcoma-carcinoma of the mammary gland, a
malignant adenoma of the liver, and sarcoma metastasis in the
kidney) with male 360, proved heterozygous to lung and medias-
tinal tumors. Note especially charts 1 and 2 where the strains
were derived from the same original mating of female 3 with male
360; and charts 4, 5, 6, 7, 8, and 9, showing ancestry and strains
derived from the mating of female 3 with male 30.

It should not be a matter for surprise that there is an inheri-
tance of a specific type of liver tissue, or kidney tissue, etc., which
will tend to respond in a given way to a given irritation or a given
lesion. Similar facts of inheritance, such as a tendency to grow
tall, with the proper degree of nourishment, etc., are easily
accepted. That also is the inheritance of tissues which tend to
respond in a given way to a given stimulus. There is here also a
certain specificity, as, for example, the tendency to longness in
the trunk and shortness of the limbs; or of shortness of the
trunk and longness in the limbs, so that a given figure tends to
predominate in a given family, just as given types and locations
of neoplasms tend to predominate in a given family. Again, it is
an accepted expectation that there should be a given nose shape,
or shape of lips, or contour of head, inherited within a family;
i.e., a specificity of nose tissue, etc., which will insure its growing
to a given shape, etc.

HEREDITY IN DETERMINING TUMOR METASTASES 171

This specificity of organ tissue, which shall insure its reacting in
a given neoplastic way, to a given chronic irritation, is exactly what
we find to be the nature of cancer heredity; and it is obvious that it
will make no difference in the nature of the reaction of such
specific organ tissue, whether the lesion is primary or secondary.

Heredity, therefore, of a specific type of organ tissue, is here
shown to be the fundamental influence in determining the inci-
dence and location of metastatic neoplasms, as well as those of
primary neoplasms.

In regard to the relative infrequency of secondary tumors in
this stock, the biological evidence is as follows: Many of these
strains have been made by hybridizing a tumorous individual
with a proved non-tumerous individual. This, of course, is
done to test for the Mendelian behavior of cancer in heredity.
Now, with the tendency to carcinoma, sarcoma, etc., segregating
out, and a specific type of organ tissue segregating out and being
transmitted as such, it is evident that any given mouse may
inherit a tendency to only one type of tumor, and in only one
organ. Other organs, therefore, in such a mouse, refuse to
yield to cancer, even where emboli are present in such organs.
Many such cases have been pointed out as shown in the charts
in this report.

Individuals into whose ancestry (as is the case in much human
heredity) an indiscriminate amount of cancer had been bred
would show a correspondingly higher percentage of metastasis.

SUMMARY

These studies in the metastasis behavior of spontaneous
tumors demonstrate the following facts:

1. In any given strain, the metastatic tumors (where there
are any) tend to occur in exactly the same organs in which the
primary tumors of that strain occur.

2. In certain strains, there is a tendency for tumors to metasta-
size in certain organs; whereas in other strains, tumors of the
same type in the same organ, even where they are of older and of
larger growth, fail to metastasize into those organs.

172

MAUD SLYE

3. Leukemia and pseudoleukemia, occurring in tumor strains,
pick out predominantly the same organs for infiltration which
show the primary and secondary tumors of that strain.

4. Tumors do not even invade by extension the organs from
which primary and secondary neoplasms have been eliminted
by heredity.

5. Individuals with secondary tumors in any given organ,
seem to be as potent as individuals with primary tumors in the
same organ, to transmit by heredity, primary tumors in that
organ.

Therefore :

6. Heredity is a strong factor in determining not only where
the primary tumors of a strain shall occur, but also where the
secondary tumors shall occur.

7. Heredity is a strong factor in determining what organs of a
strain shall yield to the invasion of leukemia and pseudoleukemia.

8. The thing which is transmitted in the heredity of oancer is
the tendency of an organ or organs to yield to cancer. This
tendency is manifested whether the lesion is primary in that
organ, or whether cells from the primary growth lodge in that
organ and form a secondary lesion.

9. The tendency to sarcoma, carcinoma, adenoma, etc., segre-
gates out and is transmitted as such.

10. A strong tendency to the location of one or more types of
tumor in a specific organ or organs, such as the liver, kidney,
pancreas, mammary gland, etc., is transmitted, owing to the
segregating out of a peculiar type of tissue in these organs, which
will respond in a neoplastic or leukemic manner to lesions of any
kind which furnish a chronic irritation of not too destructive a
type.

11. That is, there is a specificity of tissue type, from organ to
organ in a strain, which will make these organs react in a given
way to a given type of irritation.

12. It is, therefore, possible for ancestry to transmit to its
posterity every possible combination of the neoplastic or leu-
kemic tendencies which they carry either actually or potentially.

HEREDITY IN DETERMINING TUMOR METASTASES 173

13. This specificity of tissue type in organs, which will insure
its reacting in a given neoplastic manner or accepting leukemic
deposits, in response to a given chronic irritation, is what we find
to be the nature of cancer heredity, and it is obvious that it will
make no difference in the nature of the response of such specific
organ tissue, whether the lesion is primary or secondary in that
organ or organs.

Heredity, therefore, of a specific type of organ tissue, is here
shown to be the fundamental influence in determining the incidence
and location of secondary tumors and of leukemia and pseudoleu-
kemia, just as it is in determining, the incidence and location of
primary neoplasms.

It is also pointed out that any apparent testimony of the
frequent occurrence of secondary tumors in man in tissues where
primary tumors rarely occur 2 would be of no help here, as no.
human strain has ever been even partially analyzed, and no right
conclusions regarding heredity can be drawn except from analyzed
strains.

It is also suggested, that until stocks of animals to be used
in pathological and bacteriological experiments, have been thor-
oughly tested out as to their inherited potentialities, such
experiments will be lacking in any adequate control, since
heredity is not considered although it is tremendously potent.

REFERENCES

(1) Ewing: Neoplastic Diseases, Philadelphia, 1919, p. 86.

(2) Slye, Holmes, and Wells: J. Cancer Res., 1919, iv, 207.

(3) Slye, Holmes, and Wells: J. Cancer Res., 1917, ii, 401.

(4) Slye, Holmes, and Wells: J. Cancer Res., 1921, vi, 57.

(5) Slye, Holmes, and Wells: J. Cancer Res., 1917, ii, 14.

(6) Loeb, L: Univ. Penn. Bull., 1906, xix, 113.

(7) Lewin: Ztschr. Krebsforsch., 1908, vi, 273.

(8) Bashford, et. al: Third Sci. Report, Imperial Cancer Research Fund,

1908, 322.

(9) Woglom: J. Cancer Res., 1918, iii, 47.

(10) LeCount: Reference in J. Cancer Res., 1917, ii, 6.

(11) Ewing: Neoplastic Diseases, Philadelphia, 1919, p. 143.

PROCEEDINGS OF THE AMERICAN ASSOCIATION FOR
CANCER RESEARCH

FOURTEENTH ANNUAL MEETING

Held in Cleveland, Ohio, March 24, 1921

1. Report of the Council

The meeting of the Council was held at the Hotel Statler in Cleve-
land, Ohio, on the evening of Wednesday, March 23, 1921.

The following members were present: Dr. Robert H. Greenough,
president; Dr. Francis C. Wood, and Dr. William H. Woglom. Absent,
Dr. James B. Murphy, Dr. E. R. LeCount, Dr. Willy Meyer, and Dr.
James Ewing. As four members of the Council are required for a
quorum, business was transacted temporarily with the idea of obtaining
subsequently the vote of another councillor. Dr. Murphy's vote sus-
taining the action of the Council was subsequently received by mail.

The treasurer's report was read and accepted.

The following officers were elected to serve for the ensuing year:
Dr. James B. Murphy, president; Dr. Willy Meyer, vice-president;
Dr. William H. Woglom, secretary and treasurer (re-elected).

Dr. H. Gideon Wells was elected Councillor to succeed Dr. E. R.
LeCount, whose term of office expired.

The present Council, therefore, with the years of retirement, is as
follows:

Dr. F. C. Wood, 1922 Dr. Robert B. Greenough, 1925

The present Editorial Board was continued in office. It is composed

Dr. James B. Murphy, 1923
Dr. William H. Woglom, 1924

Dr. Willy Meyer, 1926
Dr. James Ewing, 1927

Dr. H. Gideon Wells, 1928

Editor, Dr. Woglom
Associate Editor, Dr. Prime

Dr. Bloodgood
Dr. Wells

Dr. Loeb
Dr. Ewing

Dr. Tyzzer

The following resignations were accepted:

Dr. C. L. Alsberg

Dr. D. B. Phemister...
Dr. Robert G. Leconte
Dr. F. P. Gay

Associate
Associate
Associate

Active

175

176

PROCEEDINGS

There has been one death during the year — Dr. H. H. Janeway, an
active member, who died on February 1, 1921.

The following gentlemen were elected to membership :

Active

Dr. Herbert U. Williams, Univ. of Buffalo

Dr. David Marine, Montefiore Home and Hospital

Dr. William Ophuls, San Francisco, Calif.

Dr. H. T. Karsner, Cleveland, Ohio

Dr. Wm. Bayard Long, New York

Dr. Leo Buerger, New York

Dr. James H. Wright, Mass. General Hospital

Dr. Carl V. Weller, Ann Arbor, Mich.

Dr. Georgine Luden, Mayo Clinic

Dr. Henry Albert, Univ. of Iowa

Dr. Eugene L. Opie, Washington Univ. Sch. of Med.

Associate

Dr. John G. Clark, Philadelphia, Penn.

Dr. Stuart Graves, Louisville, Ky.

Dr. Harris P. Mosher, Marblehead, Mass.

Dr. Kenneth T. Taylor, New York

Dr. G. R. Minot, Boston

Dr. Seth Milliken, New York

Dr. John L. Yates, Milwaukee, Wis.

Dr. Charles Norris, Chief Med. Examiner, New York

Dr. D. Crosby Greene, Boston

Dr. F. S. Mandlebaum, New York

Dr. Hugh H. Young, Johns Hopkins Hospital

Dr. Henrv A. Christian, Peter Bent Brigham Hospital

Dr. Otto V. Huffman, New York

Dr. L. Duncan Bulkley, New York

Dr. Joseph A. Blake, New York

Dr. Ward J. MacNeal, Forest Hills, N. Y.

Dr. Channing C. Simmons, Boston, Mass.

Dr. Otto Krehbiel, New York

Dr. Julius Rosenstirn, San Francisco, Calif.

The application of Mr. Donald C. A. Butts, who has not yet pub-
lished any articles on cancer, was laid on the table, and the application
of Dr. Rex Duncan was postponed.

SCIENTIFIC SESSION

A letter from Dr. H. J. Conn, Chairman of the Committee on Bacterio-
logical Technique, of the Society of American Bacteriologists, calling
on biologists to cooperate with that Society in helping to secure a reli-
able domestic source of stains was brought to the attention of the Asso-
ciation of Cancer Research at its scientific meeting on March 24, by
the President, Dr. Robert B. Greenough.

It was moved by Dr. Wood and seconded by Dr. Gaylord that the
Association encourage the manufacture of American dyes. The motion
was carried.

PROCEEDINGS

177

2. Repokt on the Journal of Cancer Research
Dr. F. C. Wood (New York):

SUMMARY

It has seemed to the Council a pity to allow the Journal of Cancer
Research to lapse, inasmuch as the volumes so far printed contain
most important papers on cancer research, and when a journal ceases
publication the libraries put the bound volumes on the back shelves and
no one ever sees them again, unless speciaj.lv inquired for. The Asso-
ciation has resigned all responsibility for the Journal of Cancer
Research and the Crocker Fund, Columbia University, has assumed
it. There will be no change in form or publishers at the present time.
The Crocker Fund will have to meet an annual deficit in the cost of
publication and I hope the members of the Association will do all they
can to increase our subscription list, because unless this is done the
Crocker Fund will have to continue the present practice of the Edi-
torial Board and charge pretty heavily for cuts and reprints.

If we could get an additional three or four hundred subscribers it
would greatly reduce the cost of publication to the individual writers.

discussion

Dr. Robert B. Greenough (Boston) : The Council felt that the Crocker
Fund is coming to the rescue of the Association in the matter of the
Journal of Cancer Research and that there is a distinct obligation
upon this Association that the Journal should not be abandoned; for
much of the work in the past six or seven years in this country has been
published in this Journal. If abandoned, the material in the earlier
editions is lost and not readily accessible to new investigators. It is
thus an obligation to the contributors that the Journal should be con-
tinued, and apparently only by means of assistance from the Crocker
Fund can this be done. For these reasons the offer of the Crocker
Fund was accepted.

3. Further Studies in Radiation Dosage
Dr. F. C. Wood and Dr. Frederick Prime (New York) :

SUMMARY

Dr. Wood showed a series of lantern slides illustrating the fact that,
contrary to statements emanating from Germany, there is no destructive
carcinoma or sarcoma dose ; that is, no fixed amount of x-ray can be
assumed to destroy the cells of any one tumor, for apparently the dosage
differs greatly both in man and animals for the same microscopic type
of tumor.

As the charts showed, in animal tumors the required dosage is from .
2 to 8 erythema doses, and the sarcomata vary in resistance as well

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 2

178

PROCEEDINGS

as the carcinomata. One interesting thing is the greatly delayed
appearance of tumors at a point where the cells received a sublethal
dose. In a mouse tumor which ordinarily appears within a week and
gains a fair size in two weeks, the appearance may be delayed for a
month or more and the tumor grow very slowly afterward. If a trans-
plantation be made from this tumor into another animal the growth
rate will rapidly approximate normal. Inasmuch as two or three
months in a mouse's life is equivalent to the same number of years in
a human being, the question must be raised whether we are not going
to see late recurrences, three or four or five years after the sympto-
matic cure of malignant tumors in man, when such symptomatic cure
is obtained by either radium or x-ray.

Most of our experiments have been made by raying the tumor par-
ticles in vitro and then transplanting them; if the tumor be rayed in
the mouse the animal is killed in a few days by the radiation. If the
tumor be immediately transplanted it grows in the new host, unless it
has been given a lethal dose. This lethal dose is approximately 10
to 20 per cent more than when the cells are exposed in vitro.

DISCUSSION

Dr. E. T. Bell (Minneapolis): What is your opinion of the treat-
ment of cancer of the cervix uteri with very high voltage apparatus
such as is now in use in Germany?

Dr. F. C. Wood: Replying to Dr. Bell's inquiry, whether rays from
new high voltage machines are more effective than those from the older
type of apparatus, Dr. Wood said he had not found that there is any
increase in the destructive action of rays given off by tubes running at
1S0,000 volts as compared with those running at 120,000 volts, meas-
uring peak voltages between 12.5 cm. spheres. In all his tests there
had been no question of absorption or of scattered radiation, the tumor
having been directly exposed to the rays. The advantage of the
high voltage machines lay in the greater penetration which can be
obtained, but there is no reason to assume that they would be any more
effective on superficial growths for equal x-ray dosage. Reports
received recently from German clinics seem to indicate that some of
the published results are greatly exaggerated.

In reply to an inquiry from Dr. Greenough regarding stimulation
effects, Dr. Wood said that he had not noticed any increase in growth
rate of the tumors, unless very small doses were given; and that the
tumors which had been greatly slowed at their first transplantation by
large doses, grow at their regular rate in the second transplantation.

In response to an inquiry as to the exact mechanism of the effect
of x-ray on the cell, and whether the connective tissue does not play
an important part in the destruction of the tumor cell, Dr. Wood replied

PROCEEDINGS

179

that he had not the slightest idea of the nature of the changes causing
destruction of the cells by rc-ray or radium. All we know is that the
division mechanism is interfered with, and that cells in a resting phase
are much more resistant to radiation than those in mitosis. Experience
with animals proves that it takes very large doses to kill all the cells of
a tumor; those in the center of the growth are killed partially by direct
action and partially by the thrombosis which occurs early in the smaller
capillaries. The periphery of the tumor, however, is, as a rule, better
nourished and there the cells require maximum dosage. An absolute
cure requires the killing of every cell during a single exposure, for if
divided doses be given, some cells may recover from the radiation
effects. Dr. Wood had not observed that connective tissue is capable
of destroying cancer cells. It is perfectly possible that .such cells might
be inclosed in dense scar tissue and remain quiescent for a considerable
period, but there is a possibility that they might again begin to grow
years after their inclusion. This had been frequently observed in human
tumors in the late recurrences in operation scars. Dr. Wood had seen
some such recurrences five, six, seven, or eight years after operation.

In response to an inquiry as to the voltage used, Dr. Wood said that
this was 120,000 volts, peak, measured between 12.5 cm. spheres. The
tumor fragments used in vitro were about 1 to 2 mm. in diameter. If
whole tumors be used, or if the tumor be exposed in the mouse, the
superficial layers nearest the tube will be killed, while the deeper layers
are not. At this voltage, 10 cm. of tissue absorbed 50 per cent of z-ray
incident on the surface; this 50 per cent includes not only directly
transmitted rays but also scattered ones. His own experiments had
all been devised to avoid the complications brought into the problem
by scattering and absorption, the effect of which can be easily deter-
mined by an ionization chamber; this varies in human cases with the
individual, because of the varying depth of the tumors. Dr. Wood's
dosage was obtained in a way which made it independent of such
effects.

Dr. William Duane (Boston) : The difference between the dosage
measured by Dr. Wood and that reported from Germany, may be due
to scattering. According to recent reports, at a distance of 10 cm.
below the skin scattering may increase the dosage 2 J times. If these
pieces of tumor be surrounded by other bits of tissue, they may perhaps
receive 2| times as much radiation as they do under the experimental
conditions described by Dr. Wood.

4. High Frequency X-ray Spectra
Dr. William Duane:

DISCUSSION

Dr. Wood: We all know, of course, that these experiments of Dr.
Duane and the results which he has shown us, form the foundation of

180

PROCEEDINGS

all our x-ray work. They underlie all the measurements of the amount
of the x-ray which reaches the body; and when his results on the physical
side meet mine on the biological side, we shall know the whole story.
He is investigating the methods of measurement of the quality and
quantity of x-ray ^ -inch comes through our filters, while I am trying to
determine how much of what he thus has measured is necessary to kill
a cell. In a few years we shall know the effect of long rays or short.
The one question which the practitioner has to decide is, whether the
human body will stand all the radiation necessary to kill all the cancer
cells. If it will not stand this amount, we cannot cure cancer with
radiation; if it will, we can. I recently gave a patient something like
45 erythema doses for a melanosarcoma of the foot; the blood changes
were nominal; the only effect was that the patient was somewhat nau-
seated. The foot was of necessity removed within twenty-four hours
to avoid absorption effects from the dead tissue. This test shows that
such a dose can be given on a part of the body where the rays do not
penetrate important organs; but when it comes to putting large doses
into the abdominal cavity it is a different question. The question of
distance is also important. If we could only get a tube such that we
could place the patient at a distance of a meter, conditions would be
better than at present, where we work with a distance of 30 or 40 centi-
meters. The next step is improvement in the wattage capacity of
the tube. Somebody must design a tube which will stand longer
quantities of current than we can now employ. The amount of energy
is very small, and exposure, therefore, amounts to six or eight hours,
which is too much for a patient to stand. If we cannot contrive to
shorten the exposure by use of a very powerful tube, we shall not get
much farther ahead than we are at the present time.

Dr. Duane: The x-ray energy increases as the square of the voltage.
When the voltage is doubled, four times as much radiation is obtained,
thus decreasing the length of time required for exposure.

5. The Occurrence of Squamous-Cell Carcinoma in Mice

Miss Maud Slye, Miss Harriet F. Holmes, and Dr. H. Gideon Wells
(Chicago): Presented by Dr. H. Gideon Wells:

SUMMARY

Review of the literature on the comparative pathology of carcinoma
shows that, in general, squamous carcinomas are not so common in
other animals as in man, with the possible exception of the dog, in which
carcinoma of the skin is not infrequent. This infrequency in animals is
probably more real than statistics indicate, since the skin carcinoma
will usually be recognized, whereas the great majority of internal
growths will escape attention. In horses, carcinoma of squamous
character is seen occasionally about the external genitalia of both sexes.

PROCEEDINGS

181

Fowls and other birds not infrequently show carcinoma of the skin, and
such growths have also been observed in fish and amphibia. A few
tumors of this group have been described in rats^ and squamous-cell
tumors of the mouth and skin have been observed in mice as well as a
few cases of squamous-cell carcinoma arising from the cardiac end of
the stomach. Most of the squamous-cell carcinomas described in mice
have occurred in the mammary gland, apparently arising either from
the ampullae of the lacteal ducts or by metaplasia of a columnar
cell carcinoma.

In 28,000 consecutive autopsies performed on mice of the Slye stock,
which had been permitted to live as long as possible without any
experimental manipulations whatever, we have observed the following
instances of primary neoplasms of squamous or stratified epithelial
structure.

Primary squamous-cell carcinoma of the skin and mouth 71

Primary basal-cell carcinoma of the skin 15

Primary keratinizing cell carcinoma of the mammary gland 56

Primary squamous-cell carcinoma of the stomach 4

Primary keratinizing cell carcinoma of the lung 1

Primary squamous-cell carcinoma of the rectum 2

Primary squamous-cell carcinoma of the vagina 1

Primary stratified carcinoma of meibomian gland 2

Primary sebaceous gland adenocarcinoma 1

153

The squamous-cell carcinomas of mice are distinguished especially
by the infrequency of metastasis and they generally show a relatively
slight tendency to infiltrate, although we have observed even infiltra-
tion of the skull and spinal column. This probably depends on the
fact that secondary infections usually kill the mice at an early stage
of the disease. All the basal-cell carcinomas and the great majority
of squamous carcinomas arose about the head, neck, and mouth, these
being the sites of the greatest amount of irritation. Frequently they
have arisen in healed wounds.

(A full report of this material will be published in a forthcoming
number of the Journal of Cancer Research.)

discussion

Dr. Wells: In reply to a question regarding the life span of the
mouse, Dr. Wells replied that it depends upon the stock. Miss Slye
has a strain of Japanese waltzing mice which become senile before they
are a year old. The chief cause of death in old mice is chronic nephritis.

Miss Slye: The average published life of the mouse is years, but
many mice in my laboratory live to be six years old or more; the average
age, however, is from three to four years. Every effort is made to keep
them alive as long as possible.

182

PROCEEDINGS

Dr. Wells: I may state that these squamous-cell carcinomas in mice
arise distinctly at a later age on the average than the other carcinomas.
This is true in human squamous-cell carcinomas, too.

Dr. William H. Woglom (New York) : In a paper published in the
Journal of Cancer Research I have discussed the frequency of
keratin in mammary carcinoma of the mouse, and have estimated it at
about 23 per cent.

6. The Influence of Heredity in Determining Tumor Metastasis:
Studies in the Incidence and Inheritability of Spontaneous
Tumors in Mice. Fifteenth Report

Miss Maud Slye:

SUMMARY

Metastases in this stock are somewhat rare. In 29,000 autopsies,
furnishing something over 4000 primary spontaneous tumors, about
19 per cent of the growths metastasized. The general testimony of
those who have discussed metastasis at all, seems to be that the occur-
rence and location of metastasis is partly a mechanical matter, and
partly a striking specificity of localization of secondary growths. This
specificity they make no attempt to explain.

Ewing, speaking of metastasis in human tumors, states that in
highly vascular t issues like the lip, stomach, and testicle, very small car-
cinomas may yield distant metastasis. But these very organs are among
those whose tumors rarely metastasize in this stock, and then only locally.
For example, of our carcinomas of the lip none has metastasized; of
growths in the testis, one metastasized locally; while among carcinomas
of the stomach three metastasized into the regional lymph-nodes only.
And this in spite of the close resemblance between these neoplasms and
human tumors in similar organs.

In this stock mammary gland tumors rarely metastasize into the
regional lymph-nodes, in contrast to those of the human breast. In-
deed, in many cases in this stock, carcinoma and sarcoma of the mammary
gland can be seen growing up to the lymph-node but not invading it.
On the other hand, pulmonary metastases from mammary gland car-
cinoma and sarcoma are common in this stock.

These divergences in the metastatic behavior of tumors of similar
type and in similar organs, require some explanation other than a mere
mechanical tendency of certain types of tumor to metastasize in certain
locations.

Metastasis in this stock has been under observation for ten years and
the data here given are based on over 29,000 autopsies.

Briefly stated, these results show conclusively: (1) In any given
strain the metastatic tumors tend to occur most frequently in exactly
the same organs in which the primary tumors of that strain occur.
For example, if a strain is high in primary tumors of the liver, many

PROCEEDINGS

183

secondaiy tumors will be found in that organ. Again, if a strain yields
large numbers of primary lung tumors, metastasis will tend to occur in
the lungs from primary tumors in almost any location. A strain high
in primary kidney tumors is high also in secondary kidney tumors, etc.

(2) In certain strains there is a tendency for tumors to metastasize
into certain organs; whereas in other strains, tumors of the same type
in the same organ, even where they are older and of larger growth,
fail to metastasize into these organs. For example, in certain strains
practically every carcinoma of the mammary gland metastasizes into
the lungs; I have many strains showing 100 per cent such behavior.
On the other hand, in other cancer strains exactly this same type of
mammary gland carcinoma, even where older and of larger growth,
never metastasizes into the lungs; and I have many strains showing
100 per cent of this type of behavior. Indeed, even where tumor emboli
reach the lungs in these non-pulmonary strains, the emboli fail to take
hold and no tumor is formed. This is definite proof that the lung here
fails to yield, even when tumor cells are present.

Again, strains which never yield primary lung tumors never yield
secondary lung tumor. Strain 164 is a striking example of this, one
family yielding 36 carcinomas and sarcomas of the mammary gland
without one case of lung metastasis. In this family also there has never
been one case of primary lung tumor.

(3) Individuals with secondary tumors in any given organ seem to be
as potent as individuals with primary tumors of that organ to transmit
(by heredity) primary tumors in that same organ. For example, female
3 with a primary carcinoma-sarcoma of the mammaiy gland, and sec-
ondary sarcoma in the kidney, transmitted primary kidney tumors to
the strains derived from her, one family showing as high as 37 per cent
of primary kidney tumor.

Furthermore, individuals with secondaiy lung tumors seem to be just
as potent as individuals with primary lung tumors to transmit (by
heredity) primary tumors of the lung. Thus, many strains derived
from a female with a primary mammaiy carcinoma with metastasis
in the lungs, mated with a male showing primary lung carcinoma (double
lung-cancer parentage) show 100 per cent primary lung carcinoma.

(4) Etiological meaning: The fact that both primary and secondary
tumors of a given organ tend to occur in the same strains and fail to
appear at all in other strains, indicates that heredity is a strong factor
in determining not only where the 'primary tumors shall occur but also
where the secondary tumors shall occur.

This fact apparently means that the thing which is transmitted in
the heredity of cancer is the tendency of an organ to yield to cancer,
whether the lesion is primary in that organ or whether cells from the
primary growth lodge in that organ and form a secondary lesion.
Moreover, in strains from which tumors of a certain organ have been
eliminated by heredity, even where cells from the primary neoplasm
lodge in such organ they fail to take hold and do not form a secondary
growth. It is, therefore, evident that not only primary neoplasms,

184

PROCEEDINGS

but secondaiy tumors also, are determined in their occurrence and
their location by heredity.

DISCUSSION

Dr. G. H. A. Clowes (Indianapolis): Has anybody ever attempted
to determine whether any of these strains are more or less sensitive to
absence of vitamines?

Miss Slye: I do not think so.

Dr. Clowes: I should think that it might be a very important factor
in proliferation.

Miss Slye: I have been carrying on some dietary experiments in my
own laboratory, but they are at too early a stage for me to give any
data.

Dr. Wells: Dr. Wood who has investigated the development of
metastases in rats subjected to exploratory incision, in order to deter-
mine the rate of metastasis after such incision, has found distinct varia-
tions among control rats of different strains, in respect to metastases.
These experiments, however, were carried out with grafted tumors,
whereas Miss Slye's work deals entirely with spontaneous growths.

Dr. Greenough: As I understand it, all the matings were made before
it was known that female No. 3 had any tumor. Was her tumor dis-
covered at autopsy?

Miss Slye: Female 3 had offspring both before and after the appear-
ance of her tumor. Her mammary tumor was evident some months
before her death, while the growths of liver and kidney were found at
autopsy. Many of these matings are made before the tumors appear,
for if one waits for the appearance of a tumor one is apt to lose the
chance of offspring in a large number of cases. Mice of high cancer
ancestry should be mated early. In the majority of my cancer mice
I have records of offspring both before and after the appearance of
tumors. In the matter of tumor inheritance it makes no difference
whether the offspring precede by many months the appearance of the
growth or are born after its appearance. The grandchild or great-
grandchild may develop a tumor before the grandmother from whom
the tumor is inherited. In tumors of the liver and kidney it is difficult
to make a clinical diagnosis, though this is sometimes possible. Tumors
of the lung are frequently accompanied by a peculiar condition of the
eye, from which it is often possible to make a clinical diagnosis.

Dr. Clowes: Is there any relation between age and the rapidity at
which tumors tend to develop and metastasize?

PROCEEDINGS

185

Miss Slye: The relation of age and tumor growth has not yet been
completely worked out, but in general it is true that tumors grow more
slowly in old mice, as they do in old men. Also it may be said in
general that the tumors of old mice never grow to any great size, and
that old mice seldom are able to support more than one tumor.

7. Proofs of the Constitutional Nature of Gancer
Dr. L. Duncan Bulkley (New York) :

SUMMARY

Laboratory
findings

(Cancer not parasitic
Negative \ Cancer not contagious
[No cause for cancer

[Cancer cell an altered normal cell
Positive | Feeding experiments showing control
( of cancer growth

II.

Statistical
evidence

Control of death statistics of cancer and tuber-
culosis, especially since 1900

Steady increase of cancer deaths under surgery,
x-rays, and radium

III Bio chemical m cancer; early and late metabolic changes

evidence I *n ^e system before and after the development
[ of the local cancerous lesion

Opinion of many celebrated surgeons during the

last 100 years, to the present time
Spontaneous cures of cancer reported
Dozens or hundreds of attested cases of benefit

or cure of cancer by numerous physicians, in

this and other countries

DISCUSSION

Dr. Bulkley: I hope for comments or objections, because if I am
wrong I want to know it.

Miss Slye: It might interest Dr. Bulkley and the Association to
know that there is a standard diet maintained in my laboratory, which
is identical for tumor strains and non-tumor strains. With autopsies
at 29,000 and living inhabitants numbering about 13,000, all fed on
exactly the same diet, the tumor strains have yielded over 4000 spon-
taneous tumors, while the non-tumor strains have never yielded a single
tumor. No meat is ever used in our laboratory.

186

PROCEEDINGS

Dr. W. T. Bovie (Boston): The mortality from tuberculosis in the
state of Massachusetts is of interest in this connection. A curve
showing the progressive decrease in mortality deviates but little from
a straight line. Its course does not appear to have been changed by
any of the activities of the medical profession. Such deviations from a
straight line as do occur seem to be correlated with influxes of foreign
people.

Dr. Wells: I am familiar with Dr. Mayo's paper on the subject of
the influence of diet on gastric cancer, and it seems to me that Dr.
Bulkley has mis-interpreted it. Dr. Mayo was discussing the influ-
ence of local conditions in causing carcinoma of the stomach, and not
the influence of diet in its constitutional relation.

8. Relation of Heterosexual Characters to Neoplasms and
Developmental Errors

Dr. Otto V. Huffman (New York): Read by title:

SUMMARY

A preliminary report on the association of heterosexual secondary
sex characters with neoplasms, especially carcinomata, in patients
observed by the speaker. In noting the occurrence of status lymphat-
icus in the clinic he observed that about ten per cent of the patients
showed some degree of heterosexual secondary sex characters, especially
in regard to the pubic hair and that on the chest and on the face, and in
regard to the width of the chest and of the hips. He found in some of
these patients such evidence of maldevelopment as a patent omphalo-
mesenteric duct and gynecomastia. This led him to take note of the
secondary sex characters of patients with neoplasms. A marked degree
of heterosexuality in the secondary sex characters is not incompatible
with normal sexual life and reproduction. Four married patients who
developed carcinoma showed slight degrees of heterosexuality but the
marriages were sterile. Another four married patients who developed
carcinoma and who had some heterosexual characters did not develop
their neoplasms until after their active sexual life, that is, after the age
of fifty. The greatest degree of heterosexuality was observed in patients
who developed carcinoma early, that is, prior to fifty years of age.
This may mean that the degree of heterosexuality has some etiological
significance. Between the 100% normal male and the 100% normal
female we have a whole series of intergrades. Several observers have
stated that these intergrades are evidently on the increase. Ordinarily
they are more susceptible than the normal to acute infectious diseases,
but sanitation and public health measures may be favoring them to
such an extent that they mature and thereby arrive at the cancer age.
From the author's observations it would seem that early cancer is rela-
tively more prevalent among them than among the normal, but he is
not yet prepared to publish comparative tables that would establish
this point conclusively.

PROCEEDINGS

187

9. End-Result Reports of Cancer Operations
Dr. Robert B. Greenough and Dr. Charming C. Simmons (Boston):

SUMMARY

While many reports of the end-results of operations for the cure of
cancer have been published, and while certain standards such as the
arbitrary three or five year limit of time are generally accepted, no
uniform plan for the reporting of surgical statistics exists, so that it is
practically impossible to compare the results from one clinic with
those of another. This matter first attracted our attention in 1907,
when we reported the end-results of operations for cancer of the breast
at the Massachusetts General Hospital, and again in 1908, when we
made a study of the end-results of operations of cancer of the mouth,
tongue, and jaw. In the latter report the following statements were
made:

A comparison of these statistics with those of other writers was made and
a large number of statistical papers were consulted. It was found, however,
that the varying conditions under which the statistics were prepared made a
comparison difficult without injustice to one or another writer. It would seem
that a generally accepted standard form for the report of end-results of operations
for cancer should be secured.

In a recent paper on cancer of the breast an arbitrary standard was adopted,
and that standard has been maintained in the present communication. It is,
briefly, as follows:

1. A definite period of time has been selected ending at least three years prior
to the report, and all of the cases entered in the hospital records under the given
diagnosis have been investigated.

2. No case has been accepted as cancer without proof by pathological exam-
ination, or subsequent recurrence, or autopsy.

3. Cases which have survived, at last report, only a portion of the necessary
three-year period, are eliminated as inconclusive.

4. Cases not traced at all after discharge from the hospital, and not appearing
in the mortality statistics of their place of residence, are eliminated as incon-
clusive.

5. All cases fulfilling the above requirements are published and counted in
the statistics with such subdivision into radical and palliative operations as may
seem expedient.

It is the hope of the writers that this standard requirement may be adopted
by others for similar reports. A writer who counts as successful cases those which
have survived a period of less than the arbitrary three-year standard, or who
accepts the clinical diagnosis of cancer without microscopic examination of the
specimen, cannot in justice ask to have his statistics accepted for comparison
with those of surgeons who exercise a more careful scrutiny. It is well said by
Halsted: "It is especially true of breast cancer that the surgeon interested in
furnishing the best statistics may in perfectly honorable ways provide them."

The obvious solution of this difficulty is to publish all cases which fulfill the
present conditions and allow the reader to draw his own conclusions. If one
surgeon reports 100 cases, of which 25 are palliative and 75 radical operations,
his statistics for radical operations may well be expected to be inferior to those
of the surgeon who performs 50 palliative and 50 radical operations in the same
total number of cases. Without these figures, however, the opportunity for
comparison is lost, and the two sets of statistics apparently meet on equal terms.

188

PROCEEDINGS

Since the method above described has proved satisfactory in opera-
tion, the writers would again present it as a standard suitable for adop-
tion by the profession at large; and as a further example of its value
they would here report, briefly, a second series of cases of cancer of the
breast which were operated upon at the Massachusetts General Hospi-
tal between the dates of August 1, 1911, and April 1, 1914. During
this period the policy of special assignments of cases to different mem-
bers of the staff was in operation, and to the writers were given all
cases of diseases of the breast which entered the hospital during that
time. By following the same plan of reporting adopted in the earlier
series a direct comparison of the work in the two series of cases is made
possible. The only discrepancy lies in the fact that a five year period
of time was adopted in the second series while the three year period
was employed, as was customary at that time, in the earlier cases.

The scheme for the reporting may be summarized as follows:

A. Record all cases entering surgical wards with the specified diagnosis during

period selected.

B. Eliminate all re-entries. (No single case should appear twice in the report).

C. Eliminate all cases recurrent after previous operation in hospital or elsewhere ;

these are not cases of primary attempt to cure.

D. Deducting B+C from A we have the number of cases of cancer available
for study of operability, mortality, and other operative statistics. These
cases may then be subdivided as follows:

E. Cases of radical operation.

F. Cases of palliative operation.

G. No operation advised or performed.

H. Operative deaths.

7. Operative mortality H-t-E+F.

/. Operability (radical operations) E-s-D.

K. Operability (all operations) E-f-F-r-D.

For the study of end-results of treatment certain cases included in D are

of no value and should be deducted, viz:
L. Cases not proved to be cancer either by pathological examination of tissue,

or recurrence, or autopsy.
M. Cases untraced after leaving hospital for required interval of time — three

years, five years.

N. Cases that have died of other diseases within the required interval of time,

and without evidence of recurrence.
0. The cases remaining after deducting L, M, and N, from D, are available for

study of end-results, as follows:
P. Radical operations.
Q. Palliative operations.
R. No operation.

S. Number of cases alive without recurrence (3 years, 5 years).

T. Number of cases died (over 3 years or over 5 years) without recurrence.

U. Number of 3 year or 5 year "cures": all operations.

V. Number of 3 year or 5 year "cures": radical operations.

W. Percentage of 3 year or 5 year "cures": all operations (U^-P+Q).

X". Percentage of 3 year or 5 year "cures": radical operations (V-f-P).

End-results: Carcinoma of the breast

A. Total entries. Carcinoma Breast. . .

B. Re-entries (entered more than once)

C. Recurrence from previous operation.

I8.04-WO4
613
80
65

1911-1914

115
8
4

PROCEEDINGS

189

468

103

360

3.6%

77%

72%

89%

91%

428

19%

27%

21%

32%

D. Cases available for study of operability, mor-

tality, etc

E. Radical operation

F. Palliative operation

G. No operation

H. Operative deaths

I. Operative mortality (H^E+F)

J. Operability: Radical operations (E-rD)

K. Operability: All operations (E+F-f-D)

L. Inconclusive cases; Lack pathological exam-
ination

M. Inconclusive cases: Untraced

N. Inconclusive cases: Died within time limit . . .

O. Cases available for end-result data

P. Radical operations

Q. Palliative operations

R. No operation

S. No. cases alive and well

T. No. cases died without recurrence

U. No. 5 year "cures" (all operations)

V. No. 5 year "cures" (radical operations)

W. Percentage of "cures" (all operations)

(U-hP+Q) .

X. Percentage of "cures" (radical operations)

(V-s-P)

This paper appears in full in the Boston Med. & Surg. Jour. 1921, clxxxv, 253.

DISCUSSION

Dr. Bell (Minneapolis) : Very important considerations in the cura-
bility of carcinoma of the breast are the size of the tumor, regardless of
its histologic type, and the presence of metastases in the axillary nodes
at the time of the operation. Has Dr. Simmons any information on
the cured cases relative to these points?

Dr. Simmons: We divided the cases clinically into four groups. In
group 1 were placed cases having a small tumor and no palpable axillary
nodes. Seventy-one per cent of these cases are living without recur-
rence. Group 2 comprised slightly more advanced cases with small
palpable axillary nodes. Thirty-three per cent of the cases in this
group are living. In group 3 were placed the advanced cases, with many
axillary nodes, in which an attempt at cure by radical operation was
made. Ten per cent of the cases in this group are living. Group 4
consisted of advanced cases in which a palliative operation only was
attempted. Five per cent of this group, one case, are living.

There were two cases of colloid carcinoma in both of which there
were axillary metastases. These cases are both living without recurrence.

Infection of the axillary nodes, as shown by the microscopic examina-
tion, had a distinct bearing on the prognosis of the case. Thus there
were twenty-four per cent cures in the cases showing axillary infection,
and fifty-six per cent cures in those where the nodes were not involved.

190

PROCEEDINGS

It is interesting to note that in sixty-nine per cent of the cases the
radical operation did all that could be expected of it; that is, if the
patient died she died from remote metastases and without recurrence
in the scar or axilla.

10. Radium ix Cancer of the Bladder
Dr. George Gilbert Smith (Boston) :

SUMMARY

A report of experiences at the Huntington Memorial Hospital,
Boston, with the use of radium in 24 cases of cancer of the bladder.
Of these cases only 1 might have been suitable for radical removal.
Seven cases were quoted to show the effect of single applications of
radium in the effort to determine a standard dosage. Fifteen cases
were treated by the introduction of screened radium emanation into
the bladder cavity. Of these 9 cases had a total treatment of 600 mc.
hours or more. Analysis of these cases showed that while 2 of them
gave a temporary diminution in the amount of tumor and 1 possibly
showed a total destruction of carcinoma, none was completely cured
even clinically. Six cases were treated by the introduction of bare
emanation tubes into the tumor itself. In 3 cases this was done without
opening the bladder; in 3 it was done by cystotomy. Of these cases 1
appeared to be completely cured after 1 year. Another case dying six
weeks after operation showed no carcinoma at autopsy. With the
other cases insufficient time has elapsed to determine the result of the
treatment.

Conclusions

1. It is useless to attempt to cure with radium infiltrating carcino-
mata, which involve large portions of the bladder wall. Necrosis of
the bladder will be brought about by any dosage which will materially
influence the tumor.

2. Certain superficial cancers of the bladder may be reduced in extent
by the application of screened radium emanation to their sin-face. This
may occur without necrosis of the bladder wall.

3. To accomplish this effect, 400 mc. hours, with screening of 0.5 mm.
silver, applied not oftener than once in six weeks, has been successful,
and has not caused any considerable reaction in the bladder.

4. The greatest effect is produced by the first 3 or 4 applications of
radium.

5. If the tumor begins to grow again, further radium applications
have little deterrent effect.

6. The best way to employ radium in cancer of the bladder is by the
implantation of bare emanation tubes in the tumor, allowing one tube
to each cubic centimeter. Steel needles containing radium may be
employed in the same way, except that they must be withdrawn after
adequate exposure has been made.

PROCEEDINGS

191

7. The necrosis caused by the implantation of radium in bladder
tumors persists for at least three months.

DISCUSSION

Dr. Wood: I think that these papers by Dr. Smith and Dr. Simmons
are admirable. It is important that we laboratory men check up our
results on human beings; we care nothing for rats and mice in themselves.
One interesting point is the statement that a radiated tumor is resistant
after scar tissue has been formed. The observation has been made by
others also. I can not think that the carcinoma cell undergoes any
change, simply because of the presence of scar tissue. The radium
dosage for cancer cells in tissues not highly vascular, or in contact with
normal tissue, is very much higher than for the same cancer cells where
thrombosis can take place and cause extensive destruction. For mouse
cancer cells, the dose is 2400 mc. hours at a distance of two centimeters;
you see how resistant the cell is. Now Dr. Simmons has shown that
630 mc. hours are all that can be allowed, so that we are still far from
able to guarantee killing the cell, or from promising that other cells will
in any way affect the life of the cancer cell. I do not believe, for exam-
ple, that lymphocytes or connective tissue cells destroy cancer cells;
there is not the slightest evidence to that effect. Indeed, the cancer
cell may lie quiescent in the connective tissue for seven, ten, or fifteen
years.

11. Problems in Cancer Research
Dr. Montrose Burrows (Saint Louis) :

(Published in the Journal of Cancer Research, 1921, VI, 131)

DISCUSSION

Dr. Burrows: In reference to the discussion on the alkalinity of the
blood in cancer patients, I wish to state that we have found no direct cor-
relation between the alkalosis in many cancer patients and the anaemia
which is invariably present. Alkalosis in anaemic patients is a well
established fact. So far our conclusions in relation to anaemia have
been drawn from a comparison of the blood tests for pH which we have
had made and the blood picture recorded on the hospital history. We
are now anxious to make all these tests for anaemia ourselves. We
also feel sure that other conditions such as lesions of the pancreas and
kidneys will alter materially the pH determinations of the blood in
the Cancer patients. Certain of our cases as well as others cited in the
literature have indicated this fact.

In referring to Dr. Clowes's statement about a lipoid membrane, I
wish to call his attention to the fact that such exists in the tissue culture.
We noted this fact in the literature in an article on "The cultivation of
bladder and prostatic tumors" (Jour. Urol., vol. 1, no. 1). In analyzing
the property of stereotropism introduced for body cells by Harrison,

192

PROCEEDINGS

we found that these cells were not in contact with the cover glass or
free surface of the medium as Harrison stated, but separated from the
surfaces by a surface film of substance which is insoluble in the medium.
This was particularly easy to demonstrate in the cultures of bladder
and prostatic carcinomata.

Returning to the question of blood alkalosis in cancer, our work has
further suggested to us that the alkalosis is due to the addition of an
alkaline substance from the tumor into the blood. The pH becomes
changed because of the lack of a very active compensating mechanism
for alkalies (Bayliss).

The demonstration of an alkalosis in the plant cancers and in the
culture of the Bacillus tumefaciens gives further confirmation of this
fact.

12. Massage and Metastasis

Dr. L. C. Knox (New York, by invitation): Presented by Dr. F. C.
Wood:

SUMMARY

The general conditions underlying the formation of metastatic
deposits from malignant tumors have long been of interest to the
surgeon as determining the scope and direction of his operative pro-
cedures and even the feasibility of operation.

In this country, although the dermatologists have always made
biopsies in doubtful cases, there has been during recent years a great
deal of discussion among surgeons as to the danger of incising tumors.
Isolated instances of apparent distribution following diagnostic inci-
sions have led to the widespread feeling that such incisions tend to dis-
tribute the tumor. This attitude, strangely enough, has existed only
in the United States; the English, French, and German surgeons do not
seem to have considered the matter as of importance.

On the other hand, the relationship of massage to the production
of tumors has excited but little interest in the minds of the practitioners
of surgery and there are but few recorded clinical observations of the
spreading of tumors by massage, though such distribution takes place
very frequently under the administration of the mechanical treatment
used by the osteopath and the chiropractor to "disperse' ' tumors.
Some of the most extraordinary instances of widespread metastatic
involvement of the entire body have been seen following massage of a
carcinoma of the breast. Several instances, in which very extensive
and early metastasis occurred after small tumors had been repeatedly
examined and rather firmly squeezed or handled by the physician mak-
ing the examination, called the writer's attention to the importance of
this phase of the metastasis question.

Though the problem is one easily attacked from an experimental
point of view, but little work has been done on it, with the exception
of experiments by Dr. E. E. Tyzzer, who some years ago observed that
massage of a mouse tumor considerably increased the number of metas-

PROCEEDINGS

193

tases. His experiments were confined to only one type of tumor* and
therefore, it seemed of great practical and scientific interest to extend
if possible, his observations to a large variety of tumors of different
histological forms so as to correlate the results obtained from animals
with those observed in human beings.

It is obvious that a small-cell tumor in which the cells are closely
related to the blood-vessels, such as the lymphosarcomata, would metas-
tasize early and extensively, while the firmer fibrosarcomata might be
expected to offer considerable resistance to the removal of tumor par-
ticles into the lymph-channels and the blood-vessels. As the lym-
phatic system in animals is not as extensive as that in man, metastasis
is most frequent by way of the blood-vessels; hence, the tumor cells
reach the lungs first, and the effect of massage would be expected to
increase very greatly the number of secondary tumors in these organs.
This was found to be the case.

Two series of experiments were carried out with the same technique,
one two years after the other. The first consisted in the inoculation of
mice and rats with six strains of carcinomata and four strains of sar-
comata. About 500 animals survived for the completion of the experi-
ment. As soon as the tumors were palpable, half of them were massaged
for thirty seconds on alternate days for about two weeks; the other half
were used for controls. After this, the tumors were removed by opera-
tion, and the animals were killed at the end of thirty days. The lungs
of these animals, as well as those of the animals which died spon-
taneously during the course of the experiment, were examined micro-
scopically, and the number of metastases and emboli was noted. In
the second experiment only one strain, a polyhedral-cell sarcoma, was
used with 50 mice. The results of the two experiments were approx-
imately the same.

It was found that in all but three strains there was a larger percentage
of lung emboli in the massaged mice than in the controls; that there
was also usually a higher percentage of actively growing metastases;
and that the total number of metastatic particles of both quiescent
emboli and growing metastases constantly increased, the variation
being from 1 per cent to 37 per cent. In this respect, little difference
was found between the polyhedral-cell sarcomata and the carcinomata.
It was found, also, that fibrosarcomata are not so apt to metasta-
size when manipulated as are the other types.

13. Further Experiments on Sensitization to Heat by Radiation
Dr. William T. Bovie (Boston):

SUMMARY

The experiments reported in this paper concern the changes which
take place between the time of radiation and the appearance of the first
visible effects.

194

PROCEEDINGS

At a previous meeting I have discussed the rate of recovery of
Paramecium from the destructive action of fluorite rays, and the sen-
sitization to heat resulting from the exposure to fluorite rays.

My interpretation of the results of these investigations was incorrect,
because through faulty methods of experimentation the effects of ozone
formed by the light were not excluded.

The experiments reported in the present communication were con-
ducted in such a manner as to exclude the effects of ozone.

A very definite sensitization to heat was demonstrated.

By increasing the temperature of the organism after the radiation,
the length of the latent period is shortened. All of the changes pro-
duced by the radiation are intensified and the total number of deaths
for a constant dosage is increased.

The effect is greater the higher the temperature and the longer the
time during which the organism is maintained at the increased tem-
perature.

The temperatures used were not high enough to affect normal
unradiated organisms, nor were the effects to be observed if the organ-
isms were subjected to the increased temperature before the exposure
to the rays.

DISCUSSION

Dr. Clowes: At what temperature do these changes start?
Dr. Bovie: Room temperature, 22°C.

Dr. Clowes: You must take into consideration variation in tem-
perature, for radiation is ultimately chemical.

Dr. Bovie: We have a thermo-couple connected with the slide, and
make certain that the temperature is back at normal before we radiate.
It would not make a great difference, however, because light reactions
are not influenced by changes in temperature.

Dr. Wood: It may interest Dr. Bovie to know that we are not able
to detect any difference in the cancer cells between heating first and
radiating afterward, or radiating first and heating afterward. The
point we used was the death point determined by animal inoculation.
This is far from what he is working with. We can observe it accurately.
We get an approximation to a logarithmic curve but whether this is
the true form of the curve is doubtful.

14. Malignant Tumors of the Thyroid
Dr. Louis B. Wilson (Rochester, Minn.) :

SUMMARY

This paper presents an analysis of the pathological data concerning
290 patients with malignant tumors of the thyroid examined in the

PROCEEDINGS

195

Mayo Clinic between January 1, 1901, and January 1, 1921. Lantern
slides illustrating, the various histologic types of tumor were shown
and the clinical course of the disease in cases of the various types dis-
cussed in connection therewith. The following is a summary of the
principal points in the paper :

1. Malignant tumors of the thyroid are much more frequent than is
generally believed. Correct clinical diagnosis is frequently missed,
(a) because they may have periods of development of from five to
fifteen years and patients are not followed up long enough after opera-
tion; and (b) because not infrequently the tumor is relatively small and
the character of metastasis is not determined, owing to the rarity of
necropsies.

2. Pathologic diagnosis is difficult owing to the great variation in
the histology of the tumor and its resemblance to that of non-malignant
processes.

3. There has been a marked failure of American surgeons to report
their cases of malignant tumors of the thyroid; this should be corrected.

4. Sufficient observations are not at hand for determining the
geographic incidence.

5. The age incidence at the date of diagnosis is greatest in the fifth
decade.

6. The distribution by sex is about one man to two women.

7. Patients usually seek medical advice on the occasion of recent
rapid growth in a long standing nodular tumor of the thyroid. Some
give histories of slow continuous growth.

8. Early thorough operation gives a fair percentage of cures. Pal-
liative operation in late cases with extensive local involvement is
warranted.

9. Pathologic diagnosis must take into account the usual develop-
ment of malignant tumors of the thyroid from proliferating embryonic
adenomas.

10. The pathologist must be thoroughly familiar with the charac-
teristics of proliferating adenomas (as first described by Langhans) in
all their stages.

11. The pathologist must be on the lookout for a possible relationship
between bizarre metastatic growths and tumors of the thyroid.

12. The pathologist, in his diagnosis for the guidance of the surgeon,
must consider the relative preponderance of proliferative and degenera-
tive processes in the tumor; but a prolif crating adenoma in a patient
of cancer age should not be considered benign unless the process of
degeneration is very extensive and thoroughly overbalances that of
proliferation.

DISCUSSION

Dr. Wells: My experience in the Chicago district corroborates Dr.
Wilson's statements; there, also, tumors of the thyroid are not uncom-
mon. It is hard to find a normal thyroid in experimental dogs in

196

PROCEEDINGS

Chicago, and I have many sent to me with carcinoma of the thyroid.
There have been but a few definite tumors of the thyroid in mice in the
Slye stock.

15. Experimental Production of Tumors
Dr. F. C. Wood:

SUMMARY

Dr. Wood reported the experimental production of tumors in the
Crocker Laboratory by Dr. Frederick Bullock and Miss M. R. Curtis.
The method employed was to feed rats with ova from cat feces. The
ova were those of the Tenia crassicollis, a frequent inhabitant of the cat
intestine.

The organism penetrates the mucous membrane of the intestine,
passes to the liver and there forms a cyst. In the wall of such cysts,
single or multiple sarcomata develop after a period which is never less
than eight months. These sarcomata are of two main types, spindle-
cell and large polyhedral-cell. They are highly malignant, metas-
tasizing throughout the animals, and are transplantable, giving a high
percentage of successful transplants at the first inoculation.

The importance of the discoveiy is the large yield of tumors and ease
of the method, in contrast with the laborious tar painting process.
One strain of animals gave approximately 50 per cent of tumors among
all rats infected.

DISCUSSION

Dr. Bell: Has Dr. Wood tried any filtration experiments?

7)/\ Wood: Some years ago large series of nitrations were carried out
on various tumors in the Crocker Laboratory but without success.

Dr. Bell: I would like to ask whether the blood shows any changes in
these very malignant sarcomata? Does it suggest a leukemia?

Dr. Wood: No changes at all.

FURTHER INVESTIGATIONS ON THE ORIGIN OF
TUMORS IN MICE

VII. TUMOR AGE AND TUMOR INCIDENCE
LEO LOEB

From the Department of Comparative Pathology of Washington University School

of Medicine, St. Louis

Received for publication November 25, 1921

1. In our former publications we have analysed the relation
between tumor age and tumor rate in the various strains of
mice (1). We found that in those strains in which the tumor
rate was high the tumors tended to appear at an earlier period
of life than in strains in which the tumor rate was lower. In
addition, we found indications that there existed a special tumor
age in certain strains of mice. In general cancer age was trans-
mitted by heredity as well as cancer rate.

The present communication differs from our preceding ones
in the following respects: (a) We analyse here our total material
connectedly, while previously we considered only certain parts,
(b) The method of computation of the relations between tumor
age and tumor rate, which we used formerly, was complicated
and made a comparison between different strains and groups of
mice difficult. We now make use of a simpler way of figuring
out the relation between tumor age and tumor rate and thus
are enabled to compare more readily the different strains, (c) We
extend our analysis in various directions and explain our results
more fully on the basis of multiple factors.

2. In order to determine the tumor age of a certain family
strain or group of strains we proceeded in the following manner.
We determined the number of mice alive in the beginning of each
of the three age periods (I age period, 7-12 months; II age
period, 13-17 months; III age period, 18 months and older.) We

197

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 3

198

LEO LOEB

then determined the number of mice which developed tumors
in each of these age periods and figured out what percentage
of mice were in each age period affected by cancer. These per-
centage figures we then reduced to a basis of 100 per cent cancer
incidence, in order to obtain figures for the cancer age in dif-
ferent strains which were independent of the absolute number of
tumor mice in each strain. We may give as an example of this
mode of determination the figures obtained for the total of all
8| + 328 strains; the last figure on each line indicates the num-
ber of mice in which tumors developed in each age period in a
unit of 100 tumor mice. The figure 56.4 per cent indicates the
tumor incidence of the whole strain.

81+328

I age period. 566 mice 194 tumor mice = 34% 27%

II age period. 208 mice 117 tumor mice = 56% 44% 56.4%

III age period. 24 mice 9 tumor mice = 38% 29%

3. We divide all our strains into high, medium, and low tumor
rate strains. The high tumor rate strains have an incidence of
more than 40 per cent. The medium tumor rate strains have
an incidence varying between 20 and 40 per cent, and the low
tumor rate strains have an incidence below 20 per cent.

In our lists we state first the name of the strain, then we give
the number of mice and tumor rate in each strain and in the
last three rows the tumor incidence in each age period, reduced
to a unit of 100 tumor mice. Added in brackets to these figures
are in certain strains the absolute percentage figures of tumors
in each age period. Wherever the name of the strain as a whole
is in brackets, the figures were merely given in the list; they were
for obvious reasons not used in figuring out the incidence in
each age period in all of the strains belonging to one of the classes
(high, medium, low rate tumor mice).

If we compare the figures giving the percentage of mice affected
by cancer in the three age periods, we may draw the following
conclusions: (1) If the number of mice in a group is small, for
instance, below 50, or even somewhat higher, the percentage
age figures are no longer of value; if the number of mice exceeds
one hundred, the percentage figures begin to be much more trust-

ORIGIN OF TUMORS IN MICE 199

High rate tumor mice

STRAIN*

NUM-
BER OF
MICE

TUMOR
RATE

PERCE

I age period

STAGE OF TUMC

II age period

>RS IN

III age period

per cent

(London Blue and White)

55.0

18.0(26.0)

37.0(54.0)

45.0 (66.0)

Total 8| + 328

566

56.4

27.0(34.0)

44.0(56.0)

29.0(38.0)

(8 J + 328 JO

118

51.0

29.0(34.0)

42.6(50.0)

28.4(33.3)

(8£ + 328B)

226

56.0

25.4(32.0)

41.0(52.5)

33.6 (42.0)

[8£ + 328 (new cross)]

146

57.5

26.5

44.0

29.5

[782a (= 8| + 328)]

60.0

35.0(38.0)

65.0(70.0)

0.0 (0)

(English)

46.0

25.0(29.0)

32.0(37.0)

43.0(50.0)

(English A)

175

63.0

20.0(29.0)

45.0(64.0)

35.0(50.0)

[Family 101 (English)]

71.0

21.0(45.5)

32.0(69.5)

47.0(100)

(English Sable A)

252

70.0

28.0(46.0)

34.0(54.5)

38.0(62.0)

(English Sable B)

67.0

20.0(27.0)

49.0(65.0)

31.0(40.0)

[Family 437 (English Sable)].. .

82.0

37.0(58.0)

63.0(100)

0.0

(Smaller English Family)

56.0

46.0(42.0)

54.0(50.0)

0.0

[344 + 328 and (344 + 328) +

437] English Sable

79.0

46.0

54.0

0.0

Total English

689

67.6

25.0(40.6)

39.0(62.5)

36.0 (.54.5)

[(European 151 + Id of No. 10)

+ 101 (English) Total]

42.0

25.0(26.0)

36.0(37.0)

39.0(40.0)

[(European 151 + Id of No. 10)

+ 101 English A]

55.0

40.0

60.0

0.0

(English B)

31.0

17.7

24.0

58.3

Total Michigan Wild +

English 101

48.0

21.0(23.0)

28.0(31.0)

51.0(57.0)

(English A)

58.0

19.0

29.0

52.0

(English B)

25.0

29.5

21.5

49.0

|European 151 + Id of No. 10

254

72.0

15.0(25.0)

33.0(55.0)

52.0(88.0)

Total European 151 + lid of

No. 10 (November 8)

187

54.5

15.0

36.0

49.0

(Total European A)

65.0

14.0

36.0

50.0

(Total European B)

44.0

17.5

33.5

49.0

(8* + lid of No. 10) A

49.0

9.0

35.0

56.0

English Sable + [(European

I ~i AO \ 1 TTT1 C X"»" /si

+ 103) Fi + Hid of No. 10]

83.0

0.0

38.0

62.0

European + English Tan

(daughter of tumor mouse

146 B)

42.0

4.6

57.5

37.9

(Silver + 10 B)

64.0

17.0

20.0

63.0

Total English Tan + German.

46.6

24.0

41.0

35.0

(English Tan + German A) . . .

49.0

21.0

35.0

44.0

(English Tan + German B). . .

43.0

33.0

67.0

0.0

200

LEO LOEB

NUM-

PERCENTAGE OF TUMORS IN

BER OF

TUMOR
RATE

MICE

I age period

II age period

III age period

per cent

Unknown (probably English)

72.0

30.0

36.0

34.0

Total 198 English + (8 +

217

54.0

20.0(28.0)

41.0(58.0)

39.0(55.0)

|198 English + (8 + German

A)]

63.0

31.0

69.0

0.0

|198 English + (8 + German

B)]

135

49.0

18.5

35.0

46.5

50.0

12.0

36.0

52.0

(8 + German A)

244

41.0

13.0

40.0

47.0

IFamily 240 of (8 + German)]

43.0

29.0

33.0

0.0

Combined (Silver + English A)

and Silver + (English Sable

57.5

23.0

57.0

20.0

English Sable 344 + German . .

64.5

20.5

26.0

53.5

English Sable + (English

Silver + Id of No. 10)

65.0

23.0

77.0

0.0

English Sable + (European -f

ld No. 10)

125

55.0

25.0

40.0

35.0

(English Sable + Cream Y) . .

53.0

14.0

46.0

40.0

693 (= English Tan + Cream)

+ 773 (= 8§ + 328)

60.0

38.0

62.0

0.0

794 (German + Carter) + (8 +

German)

49.0

18.0

25.0

57.0

794 (German + Carter) + [198

(English) + (No. 8 + Ger-

man) F3]

61.0

21.0

26.0

53.0

Waltzer + No. 8

46.0

20.5

31.5

48.0

240 (8 + German) + Cream. . .

76.0

11.0

37.0

52.0

8£ + English Sable

61.0

40.6

59.0

0.0

Waltzer + English Orange and

65.0

36.0

64.0

0.0

worthy. (2) With this restriction as to numbers, we may con-
clude that the age distribution of tumors is at least as charac-
teristic of strains and serviceable for their distinction as the
tumor incidence; in fact in certain cases it is a finer instrument
for the characterization of strains than the tumor rate. The
tumor rate of two strains may be similar, but the tumor age
may allow a differentiation between the two strains.

ORIGIN OF TUMORS IN MICE

201

Thus if we consider especially the high tumor rate strains
we notice that the tumor rate of European 151 and I daughter
of No. 10 and of European 151 and II daughter of No. 10 on the
one hand and of 8| + 328 and English and its substrains on the
other hand do not differ very much, while the tumor age is quite
distinct.

The percentage figures for the first and third age period are
similar in the former two strains and differ very much from the
figures in the latter strains. There is even a difference between
English and 8| + 328, both very high rate tumor strains; in the
English the tumors appear somewhat later than in the 8J +
328 strain. This comes out even more clearly if we consider
the percentage figures for the third age period than those for
the first age period.

4. If we compare the tumor ages of the substrains with those
of the main strains we find that on the whole they agree provided
the figures used are not too small. This applies for instance to
the English, 8J + 328 and European + I or II daughter of
No. 10 and their substrains.

In the case of the strain " unknown' ' there was some indica-
tion that it was a substrain of the English. Not only the tumor
rate, but also the tumor age of this strain furnish confirmatory
evidence for this conclusion.

5. In crosses between English and the majority of other strains
in which the tumor incidence is high the tumor age is similar to
that of the English. 8J + 328 is a cross of the female offspring
of a particular English tumor mouse with an 8| male. Here we
have, owing to the influence of the English mother, a very early
tumor age which exceeded even the average age of the English
mice. On the whole, we find the English or a similar tumor age
in the following additional crosses: (European 151 + I daugh-
ter of No. 10) + 101 English, Michigan Wild + English 101.
(European 151+ I daughter of No. 10) + 101 English) A.
English Tan + German, 198 English + (8 + German), Silver
+ English, English Sable + German English Sable + (English
Silver + I daughter of No. 10), English Sable + (European +
I daughter of No. 10) (English Tan + Cream) + (8| + 328),

202 LEO LOEB

Medium rale tumor mice

STRAIN

NUM-
BER OF
MICE

TUMOR
RATE

PERCE

I age period

N'TAGE OP TUM

II age period

DRS IN

III age period

per cent

(Complete English-Cream Hy-

brids)

1057

24.3

20.0 (8.6)

42.0(18.0)

38.0(16.1)

English-Cream Hybrids, medi

urn rate

739

32 3

19.0(11.3)

41.0(24.4)

40.0(24.0)

(English Sable 344 + Black-

Cream)

23.0

18.5

40.0

41.5

(Black Cream + English

White)

143

39.0

19.0

47.0

34.0

(Cream + English, October,

1913)

109

23.0

12.5

31.0

56.5

(White Cream + White Eng-

lish)

29.0

36.0

26.0

38.0

[Cream-English (Descendants

of English Sable 1031)]

31.0

6.9

44.6

48.5

(English Sable + Cream Y,

high rate)

53.0

14.0

46.0

40.0

(693 + Cream)

28.5

[Total English Tan (121) +

267

31.0

22.0(12.0) ;38. 0(21.0)

40.0(22.0)

Total London

452

28.0

22.0(10.9)

45.0(22.0)

33.0(16.0)

(London A)

120

27.0

13.0 (7.5) 36.0(19.0)

51.0(27.0)

(London B)

38.0

29.0(16.0) 56.5 (32.0)

15.8 (8.0)

(London C)

197

28.0

11.0(25.5) '49.0(21.0)

25.5 (11.0)

(London 4S1)

0.0

55.0

18.0

37.0

45.0

Total European + 8F5

151

28.0

12.5 (6.0)

44.25 (21)

(European -f 8 F*A)

125

30.0

12.5

48.0

39.5

(European + 8 F5B)

23.0

9.5

26.0

64.5

Total Heitler

196

25.0

12.5(61.0)

64.0(31.0)

23 .5 (11.5)

(Heitler A)

102

27.0

8.5 (4.0)

91.5 (43.0)

0.0 (0.0)

(Heitler B)

22.3

19.0 (8.5)

39.0(18.0)

42.0(19.0)

101 (English) -f (European +

103)

152

34.0

5.5 (4.0)

39.5(29.0)

55.0(40.0)

Total European + English Tan

109

32.1

7.4 (4.0)

50.0(27.0)

42.6 (23.0)

(European + English Tan A). .

28.0

4.0

45.0

51.0

(European + English Tan B). .

42.0

4.0

57.5

37.9

(June, 1914, Family of Cream

35.0

4.6

33.5

61.9

Total No. 8

243

27.5

12.0 (7.0)

27.5(16.0)

60.6 (35.0)

(No. 8 A)

213

30.0

11.4 (7.1)

32.2(19.0)

56.4 (36.0)

N0.8A1

145

34.0

17.0(11.0)

33.0(21.5)

50.0(32.0)

(No. 8 AO

22.0

0.0 (0.0)

24.5(15.4)

75.5 (47.3)

(N0.8B)

10.0

0.0 (0.0)

100(20.0)

39.0

26.0

32.0

42.0

ORIGIN OF TUMORS IN MICE

203

Total 8£ + II daughter of No.

(8* + II daughter of No. 10

(8§ + II daughter of No. 10

(8£ + II daughter of No. 10) +

lid of No. 10

(European 151 + Id of No. 10)

+ 101 English B

Total Silver + Id of No. 10. . . .

(Silver + 10 A)

(Silver + 10 B)

Total medium rate Cream + Id

of No. 10

(Total Cream + 10 includes one

low rate cross)

(Cream + 10 A)

(Cream + 10 B)

(White Cream + Id of No. 10)
(Michigan Wild + English 101

Total German

(German A)

(German B)

Total 8 + German

(8 + German A)

(8 + German B)

(Family 240 = 8 + German). . .

German + Carter B

794 = German + Carter

English (344) + 8| F4 and F«. .
Cream + [19S (English) + (8

+ German) F4]

Black Cream + European. . . .

European + Cream

(Waltzer + White English) +

(Cream + 10)

Vermont Wild + English F3. . .
(1 Waltzer f Cream) + No. 8. .

German + No. 6 F4

Waltzer + White English and
(Waltzer + White English)

+ English

(Cream B June 1914 Family)..

NUM-
BER OF
MICE

110

82
28

35
245
234

370

498
174
47
149

20
42
20
22
373
244
129
14
113
25
155
33

104

19
373
32
52

99
40

TUMOR
RATE

PERCENTAGE OF TCMORS IN

per cent

40.0
49.0
14.0

36.0

31.0
37.0
36.0
64.0

34.0

27.5
36.0
26.0
34.0

25.0
40.5
50.0
32.0
34.0
41.0
21.0
43.0
33.5
36.0
25.0
33.5

32.0
28.0

21.0
36.3
31.0
37.0

26.0
35.0

I age period

II age period

III age period

per cent

8.5

40.0

01 .0

9.0

35.0

t\R ft

0.0

100

0.0

8.9

20.8

70.3

17.7

24.0

oo . o

17.0

27.0

OK) . Kl

17.0

26.0

57.0

17.0

20.0

63.0

10.0 (5

A \

32 .0 (1/

58 0 (31 OV

8.4

29.0

5.5

28.0

66.5

22.0

46.0

32 .0

12.0 (8

33 . 0 (20

29.5

21.5

15.5 (19

30.3(37

ka o (aa m

12.0

36.0

OA . \J

23.0

51.0

0.0

14.0 (S

42.0(24

44.0(25.0)

13.0

40.0

47 O

23.0

51 .0

26.0

47.0

53.0

0.0

22.0

31.0

47 ft

5.0

29.0

fifi ft

21.0

39.5

on c

oy .0

36.5(15

.0)

73 .5 (26

0.0 (0.0)

11.0

32.0

57.0

5.3

49.0

OO . /

18.9

81.1

0.0

11.0

30.0

59.0

28.6

46.4

25.0

11.9

24.0

64.1

29.0

28.0

43.0

4.6

33.5

61.9

204

LEO LOEB

German + Carter + (198 (English) + (No. 8 + German) F8),
8J + English Sable, Waltzer + English Orange. Waltzer No.
8 is similar. It is different, however, in a cross between Eng-
lish and Cream which approached the tumor rate of the English,
namely, English Sable + Cream Y. The tumor rate is here
55 per cent, but the tumor age is distinctly different; it is in-
termediate between that of the English and Cream.

German, 8 + German, German + Carter have a later tumor
age; the same applies to the crosses between these strains. Some-
what later even is the tumor age of (8 + German) + Cream,
owing to the influence of the Cream; yet the influence of the
8 + German causes the tumor age to be earlier than that of the
Cream. In a similar way the tumor age of the (8| + II daugh-
ter of No. 10) A is later, although in this group the tumor rate
is high.

MEDIUM RATE TUMOR MICE

6. While in the large majority of the hybrids between Cream
and English the tumor rate is intermediate (32.3 per cent) be-
tween that of the Cream and English strain, in some of them it
is low like that of the Cream, and in one of them, as we have
seen, it approaches that of the English strain. We have divided
these hybrids in two larger groups, one containing the low and
the other the medium tumor rate strains. We find that the tumor
age in both reaches almost, but not quite, that of the English
parent. Tumors appear relatively early in these groups; only
in two of these hybrid strains the tumor age is intermediate.
Again we notice that in small strains comprising only 26 or 27
individuals the determination of the tumor age is no longer
certain.

7. The result as to the effect of hybridization on tumor age is
different in another hybrid strain in which Cream was crossed
with the No. 10 (I daughter) strain. While in this case the
tumor rate is intermediate, the tumor age approaches that of
the Cream, although it does not quite reach the lateness of the
tumor age of the Cream. On the other hand, in a cross between
Silver and the same No. 10 strain, although the tumor rate is

ORIGIN OF TUMORS IN MICE

205

again intermediate — in accordance with the low tumor rate of
Silver which almost corresponds to that of the Cream — the tumor
age is decidedly different and very much resembles that of the
No. 10 strain; this is in accordance with the fact that, as far as
can be determined from the restricted number of Silver mice at
our disposal, the tumor age is in this latter strain very much
higher than that of the Cream.

This dissociation between tumor age and tumor rate is like-
wise apparent in the London strain, where the tumor age almost
reaches that of the English, while the tumor rate is very much
lower.

A similar dissociation we find in the following strains: Car-
ter (German + Carter) B, English (344) + 8|, Cream + (198
English + (8 + German) F4) and in the Waltzer and English
as well as the Waltzer and Cream hybrids. In the case of the
Cream + (198 English + (8 + German) F4) strain the number
of mice is perhaps too small to permit much consideration. The
Carter strain is in its origin related to the English; the tumor
rate in this case stands on the border between a medium and
high rate. The other strains in this class are hybrids in which
either English, Carter, or Waltzer mice enter. In these hybrids
there seems to be a tendency to an early tumor age. This would
agree with the early tumor age of the English and Carter parent
strains. In regard to the pure Waltzers, we have no statistics
as to their tumor age and tumor rate. To judge from the result
in the crosses we can only infer that their tumor age was early.

8. In a number of tumor strains with medium tumor rate
the tumor age corresponds to the rate. (11 to 15 per cent of
tumor mice in the I age period, an average of 45 to 55 per cent in
the III age period.) In this group we include: European +
No. 8, Heitler, German, 8 -j- German, No. 8, White Cream +
I daughter of No. 10, Black Cream + European, Vermont Wild
+ English F3 and German + No. 6 F4. Of these, Black Cream
+ European and Vermont Wild + English stand on the border
between a medium and late tumor age.

9. There are some strains with a medium tumor rate and a
late tumor age. To this group belong: 101 (English) + (Euro-

206

LEO LOEB

pean + 103), European + English Tan, 8| + II daughter of
No. 10, (Si + II daughter of No. 10) + lid. of No. 10, Cream +
I daughter of No. 10, and European + Cream. We notice that
the parents (European + 103), European 8§, and Cream have
the tendency to impart to hybrid strains a late tumor age. All
those parent strains have either themselves a late tumor age
or their tumor age is on the border between medium and late,
with exception of the European which, notwithstanding their
relatively early tumors, likewise seem to impart a late tumor
age to their hybrid offspring. In this strain there was perhaps
present a greater individual variability in the tendency towards
a certain tumor age. We notice furthermore that whenever the
I or II daughter of No. 10 enters into a cross with a second parent
strain with a tumor age which is late or approaches lateness, the
No. 10 component of the cross, despite its high tumor rate, does
not tend to make the tumor age earlier, because the European
+ I and II daughter of the No. 10 mice themselves have a rela-
tively late tumor age considering their high tumor rate.

10. If we compare the list of the high tumor rate strains with
that of the medium tumor rate strains, we find in the former the
tumor age on the whole much more homogeneous than in the
latter. In the high tumor rate strains the tumor age in all the
important strains with exception of strains into which the daugh-
ters of No. 10 enter is early, while in the medium tumor rate
strains the/variability as to tumor age is much greater, although
the deviations are about equally distributed on both sides of a
medium tumor age, which latter represents the average.

LOW STRAIN TUMOR MICE

11. The Cream strain is typical of the low rate tumor strains.
The difference in the tumor age between this strain and a typi-
cal high rate strain is striking; the tumors appear here much
later in life than in the high tumor rate strain. We must of
course take into consideration the fact that the smaller the
number of tumor mice, the greater is the effect on the percentage
age distribution of a small variation in the number of mice dying
in the different age periods. Therefore we should expect a

ORIGIN OF TUMORS IN MICE 207

Low rate tumor mice

NUM-

TUMOR

PERCENTAGE OF TUMORS IN

STRAIN

BER OF

RATE

MICE

I age period

II age period

HI age period

per cent

per

cent

per cent

Total English Cream low rate

hybrids

318

•5)

38.0 (3.4)

34.0 (3.0)

(344 _j_ Black Cream) Fi 4-

Crpam Twhi t,p^

0.0

(English 344 -f- Cream, new

individuals)

143

37.0

39.6

(English Sable 4444 -j- Cream)

175

36.0

34.0

Total Cream

878

■ 0)

26.0 (3.3)

66.0 (8.2)

(Total Cream X)

260

52.3 (2.3)

29.5 (1.3)

(Orpam X A)

135

• 4)

73.0 (3.8)

0.0 (0.0)

("Cream X H)

33.0 (2.0)

67.0 (4.0)

(Cream X C)

0.0

(Old Cream)

226

20.0 (1.2)

80.0 (4.3)

(Cream A, mostly black)

114

23.0 (3.7)

67.0(11.0)

fOrpflm R mostlv whit.p polor")

•1)

39.0(14.0)

58.0(21.0)

(Among Cream B, June 1914

Family)

a
0

33.5 (18.0)

61.9(33.0)

(The rest of Cream B after

deduction of June 1914 fam-

ily)

(Cream Black II)

0.0 (0.0)

87.0 (8.5)

(Cream Black III)

14.0 (3.1)

72.4(16.0)

(Cream Y)

0.0

(Total new Cream)

349

25.0 (5.6)

67.4(15.0)

Total Enronpnn

263

34.0 (5.4)

45.0 (7.4)

(European A)

113

42.0(12.0)

37.0(11.0)

(European B)

150

14.0 (1.0)

68.0 (5.0)

Total European + 102 or 103.

221

24.0 (7.5)

73.0(23.0)

(Euronpan 4- 102 nr 10*3 A ^

146

23.0 (8.0)

71.7(25.0)

(EuroDean 4- 102 or 103 VI)

30.5 (5.5)

69.5 (12.5)

(London 481)

4*3

0.0

Totfl 1 TjOTlflnr* -I- (Piirnnoon _i_

103) F3

148

25.0

69.0

[London + (European + 103)

30.0 (2.8)

57.0 (5.4)

[London + (European 4- 103)

F3B]

31.0 (4.1)

69.0 (9.0)

415 = [101 English + (Euro-

pean + 103)]

41.0(12.0)

47.0(3.5)

Total No. 8|

241

50.0(16.0)

39.0(10.0)

(No. 8| A)

158

53.0(16.0)

33.7(10.0)

(No. 8| B)

0.0

208

LEO LOEB

STRAIN

NUM-
BER OF
MICE

TUMOR
RATE

PERCEI

I age period

STAGE OF TUMC

II age period

)R3 IN

III age period

per cent

(No. 8£ C)

9.0

7.5

(2.0)

33.0

59.5 (16.0)

Total (European + 102) Fi +

8| F4

602

16.0

6.5

21.0

72.5

[(European + 102) Fi + 8|

F4 A]

473

16.0

7.0

18.0

75.0

[(European + 102) Fi + 8| F4

129

15.0

8.6

28.5

62.9

Total (European + 103) Fi +

Hid of No. 10

219

14.6

4.4

29.0

66.6

[(European -f 103) Fi + Hid

168

17.0

6.3

28.0

65.7

[(European + 103) Fi -f Hid

of No. 10 B]

13.0

0.0

10.7

89.3

[(European + 103) Fx + Hid

of No. 10 C]

0.0

127

7.0

35.0

(3.5)

65.0

0.0 (0.0)

English Silver Fawn

12.0

0.0

(0.0)

38.0(11

62.0(18.0)

London + Silver

176

12.0

8.0

24.0

68.0

(No. 8 B)

10.0

0.0

100

(8£ + Hd of No. 10 B)

14.0

0.0

100

0.0

German + Carter A

358

9.0

6.6

26.0

67.4

English Sable 344 + (European

+ English 146) (344-146). . . .

5.0

0.0

100

0.0

White Cream + (European

151 + Id of No. 10 = No. 697)

128

9.5

5.8(1.65)

17.5

(5.0)

76.7(22.0)

Cream + (Cream + European

428) Fx....,

10.0

0.0

100

0.0

German + No. 8

112

0.0

much greater variability in the age distribution in the low rate
than in high rate tumor strains. This does not, however, apply
in the case of the representative strains of this group because we
are working with a very great number of individuals in these
strains. Thus, while in the Cream strain the tumor rate is as
low as 5.9 per cent, the strain consists of 878 mice and among
them are 52 tumor mice, a sufficiently large number for our
purpose. If we consider, on the other hand, the substrains in
this group, then the number of tumor mice usually becomes so
small that the development of one or two more tumor mice in

ORIGIN OF TUMORS IN MICE

209

an earlier age period may change considerably the age distribu-
tion. Notwithstanding this fact we find in the Cream strain
that the smaller substrains behave essentially like the main
strain; we notice a low percentage of tumor mice in the first age
period and a high percentage in the third age period, with ex-
ception of the Cream X A substrain where we have to deal with
only 5 tumor mice, two of which died in the first and three in the
third age period, and even here a change in time of death in the
case of one or two individuals would have made the age curve
typical.

12. What applies in the case of the Cream strain applies also
to the other more important low rate tumor strains, as for in-
stance the European + 102 or 103 where, as a result of the some-
what higher tumor rate, we have to deal with 34 tumor mice. It
also applies in the case of London + (European + 103) F3, in
the large strain (European + 102) Fi + 8J F4, where we can
base our calculations on 97 tumor mice; furthermore, in the case
of the (European + 103) + HI daughter of No. 10, London
+ Silver, German + Carter A in which the tumor rate was only
9 per cent and the age distribution was accordingly typical for
the low rate tumor strain, while in the German + Carter B, where
the tumor rate rose, the percentage distribution in the different
age periods also changed in the typical manner.

The same age distribution we find furthermore in the White
Cream + (European 151 +1 daughter of No. 10) and some
smaller strains. In the German + 8 tumors did not develop.
Wherever the European + 102 or 103 enter a cross, they not
only depress the tumor rate, but call forth the typical age dis-
tribution characteristic of the low tumor rate strains. This
applies even in the case of the English 101 + (European + 103)
where the association with the English strain succeeded in rais-
ing the tumor rate to 34 per cent and thus in producing a medium
rate tumor strain; but the tumor age is very nearly as late as
that of the Cream.

13. There are a few groups in this class in which the tumors
appear earlier, namely, European, a small substrain 415 = 101
English + (European + 103) , No. 8 J, and English silver. Eng-

210

LEO LOEB

lish Silver is related to the English and the tumor age ap-
proaches that of the English. Notwithstanding the relatively
small number of tumor mice in this strain, this age distribution
is probably not altogether due to a coincidence; at least we may
interpret in this sense the fact that even in the hybrids in which
English Silver enters the tumor age is relatively early. In
No. 8f and in 415 = 101 + (European + 103) the tumor age
stands at the border between the early and medium types, or
rather it approaches the medium type. In the European strain
the tumors appear earlier. Here we may have to deal with
another instance in which tumor rate and tumor age are dis-
sociated; in the case of the low rate tumor mice the number of
tumor mice is, however, relatively so small that in individual
strains, except the largest ones, chance variations cannot be
excluded with certainty.

14. In the English-Cream hybrid with a low tumor rate the
tumor age is very early, similar to that of the English. The
number of mice is 318, the tumor rate is 5.7 per cent; we have
to deal with 18 tumor mice. This corresponds to the similar
high tumor age in the English-Cream hybrids with medium
tumor rate. In this group the number of mice is still greater
and our classification is based on the cancer age of 239 tumor
mice.

We may therefore conclude that in the English-Cream hybrids,
independently of the tumor rate which may vary in different cases,
the tumor age approaches that of the English. This applies
even in the case of the two individual English-Cream hybrids
with a low tumor rate.

In the case of the English-Cream hybrids we find, therefore,
a splitting in the inheritance of tumor rate and tumor age.

THE TUMOR AGE OF THE HIGH, AND LOW TUMOR RATE MEDIUM

STRAINS

15. The great difference in the tumor age of the high, medium,
and low tumor rate strains is brought out very clearly if we dis-
tribute all our strains among three great groups, according as to

ORIGIN OF TUMORS IN MICE

211

whether their tumor rate is high, medium, or low, treating each
group as a whole and determining in each of them the tumor age
of the group.
Thus we obtain the following figures:

Total of the high tumor rate strains

I age period. 2741 mice 885 tumors = 32.3% 22.3%
II age period. 1176 mice 636 tumors = 54.1% 37.3%
III age period. 225 mice 132 tumors = 58.7% 40.4%

Average tumor rate 60.3%

Total of the medium tumor rate strains

I age period. 4351 mice 394 tumors = 9.0% 14.6%
II age period. 2855 mice 635 tumors = 22.2% 36.0%
III age period. 1238 mice 374 tumors = 30.5% 49.4%

Average tumor rate 32.2%

Total of the low tumor rate strains

I age period. 3971 mice 67 tumors = 1.69% 8.5%
II age period. 2747 mice 152 tumors = 5 .53% 27.8%
III age period. 1362 mice 173 tumors - 12.7% 63.7%

Average tumor rate 9.9%

This arrangement brings out very clearly the difference in the
tumor age of the different groups. In the group composed of
the high tumor rate strains the average tumor rate is 60.3 per cent,
the tumors appear early, almost, but not quite, as early as in
the English strain. The figures for the I, II, and III age period
are I 22.3 per cent, II 37.3 per cent, III 40.4 per cent. In the
second group the average tumor rate is 32.3 per cent. The
tumor age is correspondingly intermediate. The figures for the
3 age periods are: I 14.6 per cent, II 36 per cent, III 49.4 per
cent. In the third group the average tumor rate is 9.9 per
cent. The average tumor age is approximately that of the
Cream; the figures are I 8.5 per cent, II 27.8 per cent, III 63.7
per cent. We may therefore conclude that the higher the tumor
rate in a strain, the earlier the tumors appear on the average;
the greater is the number of tumors in the first age period, and
the smaller is the number in the third age period. The differ-
ences in the first age period are, however, somewhat greater

212

LEO LOEB

than those in the third age period. The differences between the
figures for the II age period, which represent the turning point
are, as might be expected, much smaller in the three groups.

16. This relationship between tumor rate and tumor age can
be shown in still another way. We may arrange all the strains
in three different classes in accordance with their tumor age,
the first class comprising those strains having a tumor age simi-
lar to that of the high tumor rate group, the second comprising
those strains having a tumor age similar to that of the medium
tumor rate group, and the third comprising the strains with a
tumor age similar to that of the low tumor rate group.

We then find in the first class, comprising the strains with an
early tumor age, 19 strains with a high tumor rate, 10 strains
with a medium tumor rate, and 3 strains with a low tumor
rate. Almost 60 per cent of the strains belong, therefore, to the
high tumor rate strains.

In the second class, with a medium tumor age, we find 10
strains with a medium tumor rate, 2 strains with a low tumor
rate, and 5 strains with a high tumor rate. In this class 59 per
cent of the strains belong to the medium tumor rate strains.

In the third class, with a late tumor age, we find 14 strains
with a low tumor rate, 6 strains with a medium tumor rate, and
1 strain with a high tumor rate. Sixty-seven per cent of the
strains belong to the low tumor rate strains. In each class the
majority of strains have a tumor rate which corresponds to the
tumor age of that class.

ON THE PERIOD OF LIFE IN WHICH THE DEATH-RATE FROM CANCER
IS HIGHEST IN MICE

17. It is generally assumed that in man the cancer incidence
on the whole increases with increasing age. However, we find
certain variations according to the kind of cancer which is con-
sidered ; thus the death-rate from mammary and uterine cancer
reaches a somewhat earlier maximum than the death-rate from
cancer in general. The same applies to sarcoma. In mice
Murray finds the death-rate from mammary cancer to reach a
maximum at the age of thirteen to sixteen months. Murray's

ORIGIN OF TUMORS IN MICE

213

statistics differ from our own in that they are based on a much
smaller number of mice and in the method of computation he
uses. Later we shall apply the same method also to our material.

We found that the age of death from mammary cancer in
mice is not a fixed point, but varies with the cancer rate. The
cancer rate being variable in different strains, the cancer age
is likewise variable. In general we may state that the cancer
age is the earlier, the higher the cancer incidence.

18. If we consider the total figures for high, medium, and low
tumor rate strains respectively, we find that in the high tumor
rate strains the difference between the cancer rates in the third
and second age periods is very small, but the cancer rate is slightly
higher in the third age period. In mice above the age of eight-
een months the cancer rate is therefore slightly higher than in
mice between the age of twelve and eighteen months. The
maximum still is in the third age period.

In the medium tumor rate strains the difference between the
cancer rate in the second and third age periods is more marked;
the maximum is here quite definitely in the third age period.
In the low tumor rate strains the difference between the tumor
rate in the second and third age periods is still much greater;
the maximum is very decidedly in the third age period. The
lower the tumor rate, the more decidedly the maximum moves
to the third age period. The lower the tumor incidence, the
greater the maximum reached in the third age period and the
greater the difference between the tumor incidence in the second
and third age period. The difference between the tumor rates
in the first and second age periods varies less; it happens to be
least in the group of the high rate tumor strains.

19. If instead of considering the groups as a whole, we con-
sider the individual strains composing them, the difference in
the situation of the maximum tumor incidence in different strains
comes out still more clearly. Thus in the 8^ + 328, a high tumor
rate strain, the maximum is decidedly in the second age period,
while in the English, although they have a slightly higher tumor
rate, the maximum is somewhat higher than in the 8£ + 328,
although it is still in the second age period if we consider the

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 3

214

LEO LOEB

total strain; the difference between the second and third age
periods is however very small in this case and in the substrains
we find some variation as to the maximum.

The great majority of the English hybrid strains behave
like the English strain as far as the maximum of the tumor in-
cidence is concerned. Similarly to the English strain behave in
this respect particularly the English-Cream hybrids and Euro-
pean + English Tan, irrespective of their tumor rate.

In addition to various English hybrids, the strains London,
8 + Gennan, and 8£ show a maximum similar to that of the
English mice. On the other hand, in the European 151 + I
or II daughter of No. 10, in which the tumor incidence is very
high, the maximum is decidedly in the third age period.

We must then conclude that the maximum of tumor incidence
is found at a different age level in different strains, and that this
peculiarity is transmitted by heredity in the same way as the
other characteristics of the strain.

The data as to the age period in which the maximum cancer
incidence is found in mice, which we present here, permit only
comparative conclusions. We are at present operating with
three age periods which are not equal in length and we cannot
therefore state at which age period the maximum occurs. But
our data have a relative value. They show the shifting of the
age period in different strains in dependence upon the variations
in the cancer rate of the various strains and groups. And we
may furthermore conclude that in all statistics so far no account
has been taken of this variability of the tumor age in general
and of the maximum in the tumor rate in particular. All avail-
able statistics give an average, in which this variable factor is
disregarded and in which necessarily the result will differ in
accordance with the relative preponderance of high or low tumor
rate families or strains.

20. If instead of considering the three groups of mice sepa-
rately we determine the tumor incidence in the three age periods
in all the mice, irrespective of the group to which they belong,
we obtain the following figures: In the first age period there
were 11,063 mice; 1346 mice developed tumors during this period
= 12.2 per cent.

ORIGIN OF TUMORS IN MICE

215

In the second age period there were 6778 mice with 1423
tumor mice = 21 per cent.

In the third age period there were altogether 2825 mice, with
679 tumor mice = 24 per cent.

If we calculate these figures on the basis of a tumor incidence
of 100 per cent, we find for the first age period a tumor incidence
of 21 per cent, for the second age period an incidence of 36.7 per
cent, and for the third age period an incidence of 42.3 per cent.
These figures are somewhat between those of the high and medium
tumor rate mice, but they approach very closely those of the
high tumor rate mice. The maximum of the tumor incidence
is accordingly in the third age class, but the difference between
the figures for the second and third age period is relatively small.

21. If instead of considering in each age period the number,
of mice alive at the beginning of this period and the number of
mice developing tumors in this period, and on this basis de-
termining the percentage figures of tumor incidence for each
age period, we determine what percentage of mice will develop
cancer during the whole period following this date irrespective
of the number of mice eliminated through death during each
period, we obtain a different result. We had altogether 11063
mice which reached the age above six months. Of these 3448
subsequently developed cancer = 31.2 per cent; of 6778 mice
which reached the age of twelve months, 2102 later developed
cancer = 31 per cent; and of 2825 mice reaching the age of
eighteen months, 679 developed cancer = 24 per cent. Figured
out on this basis the tumor rate is similar for mice reaching the
age of six and of twelve months. In these groups the tumor
rate is higher than in mice reaching the age of eighteen months.

This mode of computation seems to us less satisfactory than
the first method used, because it includes the number of indi-
viduals dying in each period among those which are liable to be
affected by cancer in subsequent periods, and it thus makes the
tumor incidence in the first two age periods too high.

22. If we leave out of consideration the mice dying below the
age of 6 months we find among our material altogether a tumor
incidence of 31.2 per cent. This is a relatively high tumor rate

216

LEO LOEB

which presumably exceeds the average tumor rate among white
mice in unselected strains. Our material is a selected one, in
which a special value was attached to high tumor rate strains.

23. If we determine with the second method the cancer inci-
dence in the various groups of high, medium, and low tumor rate
mice, we find the following figures.

High tumor rate mice

I 2741 mice 1653 tumor mice = 60.3%
II 1176 mice 768 tumor mice = 65.3%
III 225 mice 132 tumor mice = 58.7%

Medium tumor rate mice

I 4351 mice 1403 tumor mice - 32.3%
II 2855 mice 1009 tumor mice «= 35.3%
III 1238 mice 374 tumor mice - 30.2%

Low tumor rate mice

I 3971 mice 392 tumor mice = 9.9%
II 2747 mice 325 tumor mice « 11.7%
III 1362 mice 173 tumor mice « 12.7%

On this basis the percentage figures for the different age periods
are very similar in the three groups. But in the high and
medium tumor rate mice the maximum of the tumor incidence is
in the second age period, in the low tumor rate mice it is again
in the third age period. Even this method, which is more
faulty than the first one, brings out the shifting in the tumor
incidence with varying tumor rates.

THE INHERITANCE OF TUMOR AGE AND INCIDENCE IN HYBRIDS

24. We have already referred to the behavior of individual
hybrid strains as far as the relation between tumor age and tumor
rate is concerned. We have also referred to some of the larger
groups of hybrids. We shall now discuss the behavior of hybrids
in some of the larger groups connectedly and compare their
tumor rate and age with that of the parent strains.

(a). English-Cream hybrids. The figures for the parents are
as follows:

ORIGIN OF TUMORS IN MICE

217

A. Total English

I
II
III

689 mice
258 mice
46 mice

280 tumor mice = 40.6% 25%
161 tumor mice = 62.5% 39% 67.6%
25 tumor mice = 54.5% 36%

B. Total Cream

I
II
III

878 mice
600 mice
279 mice

9 tumor mice = 1.0%
20 tumor mice = 3.3%
23 tumor mice = 8.2%

8%
26%
66%

5-9%

C. Total English-Cream Hybrids

I 1057 mice 92 tumor mice = 8.6% 20%

II 668 mice 120 tumor mice = 18.0% 42% 24.3%
III 280 mice 45 tumor mice = 16.1% 38%

In these hybrids the tumor rate is on the whole intermediate
between those of the parents, although somewhat nearer the Cream
parent than the English. The tumor age, on the other hand, is
almost that of the English parent . In this case we have to deal
with such large figures that we are justified in assuming that
these figures express causal relations and are not due to accidental
findings.

The same relation comes out still more clearly if we divide the
English-Cream hybrids into three subgroups.

A. Total English-Cream Hybrids with a medium tumor rate
(and with a slight admixture of high rate tumor mice)

I 434 mice 53 tumor mice = 12.0% 20%

II 264 mice 67 tumor mice = 25.0% 41% 32(%

III 100 mice 21 tumor mice = 23.0% 39%

B. Total English-Cream Hybrids with a low tumor rate

I 318 mice 8 tumor mice = 2.5% 28%
II 205 mice 7 tumor mice = 3.4% 38% 5.7%
III 101 mice 3 tumor mice = 3.0% 34%

In this case the tumor rate is altogether that of the Cream,
but again the tumor age is that of the English parent strain.

C. Total English Tan (121) + Cream

II
III

267 mice 32 tumor mice = 12%
173 mice 36 tumor mice = 21%
63 mice 14 tumor mice = 22%

22%
38%
40%

31%

218

LEO LOEB

Also in this strain of hybrids the tumor rate is intermediate
while the tumor age is that of the English parent strain.

b. The same conclusion, that factors for tumor rate and age
may be transmitted independently of one another, follows from
a consideration of the hybrids between Cream and European
151 + I daughter of No. 10.

The tumor rate and tumor age of the Cream we have already
given.

Figures for the mother strain European 151 + I daughter of No. 10 are as follows:

I 254 mice 63 tumor mice = 25.0% 15.0%
II 148 mice 120 tumor mice = 81 .0% 33.0% 72.0%
III 46 mice 39 tumor mice = 88.0% 52.0%

The Total Hybrids Cream + (European + I daughter of No. 10)
(three different crosses used)

I 498 mice 22 tumor mice = 4.4% 8.4%
II 365 mice 53 tumor mice = 15.0% 29.0% 27.5%
III 189 mice 62 tumor mice = 33.0% 62.6%

In this case the tumor rate is again on the whole intermediate
and slightly nearer that of the Cream, but the tumor age is that
of the Cream, contrary to what we found in the former hybrids.
If we omit from the table of hybrids one cross in which the tumor
rate was low, the figures for the Cream + 10 hybrids with medium
tumor rate (2 crosses) are as follows:

Total Cream + (European + I daughter of No. 10) hybrids with medium tumor rate

I 370 mice 20 tumor mice = 5.4% 10%
II 286 mice 49 tumor mice = 17.0% 32% 34%
III 161 mice 56 tumor mice = 31 .0% 58%

Again the tumor rate is intermediate and the tumor age is
almost that of the Cream. In this connection it is of interest
to remember that the tumor age of the European 151 + I or II
daughter of No. 10 is later than that of the English, and accord-
ingly it has a tendency to transmit to crosses a later tumor age
than the English.

If we cross strains in which tumor rate and tumor age differ
less than in the preceding strains, we again obtain a tumor rate

OKIGIN OF TUMORS IN MICE

219

and tumor age intermediate on the whole, or perhaps resembling

slightly more one parent than the other. This is shown in the
following crosses:

A. German

I 42 mice 8 tumor mice = 19% 15.5%

II 19 mice 7 tumor mice = 37% 30.3% 40.5%
III 3 mice 2 tumor mice = 66% 54.2%

In this case the number of mice is rather small; we can there-
fore use it only with caution.

B. Total No. 8

I 243 mice 17 tumor mice = 7% 12.0%
II 150 mice 24 tumor mice = 16% 27.4% 27.5%
III 78 mice 26 tumor mice = 35% 60.6%

C. Total No. 8 + German

I 373 mice 30 tumor mice = 8% 14.0%
II 280 mice 66 tumor mice = 24% 42.0% 34.0%
III 123 mice 31 tumor mice = 25% 44.0%

The tumor rate of the hybrids is about intermediate and the
tumor age is very similar in both parent strains as well as in the
hybrids.

d. Another strain may be of interest; this resembles in its
character the last named hybrid strain and is related to it. We
crossed the 8 + German with an English male, the offspring of
Tumor mouse 198.

The data for the parent strains have already been given. The
figures for the cross are as follows:

English (198) + (8 + German)

I 217 mice 61 tumor mice = 28% 20%

II 89 mice 52 tumor mice = 58% 41% 54%

III 9 mice 5 tumor mice = 55% 39%

In this case the tumor rate of the hybrids is again intermediate,
but slightly nearer that of the English. The tumor age is also
intermediate, but again somewhat nearer that of the English.
In this case no distinct split between tumor rate and tumor age
occurred in the hybrids.

220

LEO LOEB

25. We believe that in these investigations we have established
a definite relation between tumor incidence and tumor age. In
general the average tumor age is the earliest in those strains of
mice in which the number of individuals affected by cancer of the
breast is greatest. There is in addition a difference in the tumor
age in a number of strains in which the tumor rate is similar;
or, conversely, in a number of strains in which the tumor age is
similar the tumor rate may differ. We have seen furthermore
that in crosses tumor age and tumor rate may be inherited in-
dependently of each other, or in other cases the tumor age may cor-
respond to the tumor rate. How can these facts be interpreted
in accordance with the current theories of heredity? We may
assume as most probable that the tendency to cancer depends on
the presence of multiple factors.

A certain number of these factors must be present in an indi-
vidual if a tumor is to appear. These factors determine the in-
tensity of the tendency towards cancer in each individual; but this
intensity tends not only to cause a tumor to appear in a certain
animal, but it has the tendency to make it appear at an early
period of life. Tumors appear the earlier, therefore, the greater
the number of tumor factors present or the greater the impor-
tance of the factors represented. We may assume that different
strains differ greatly in the number and character of these fac-
tors present in the average of the individuals. In the English
and 8£- + 328 there may be in an individual affected with can-
cer, on the average, a greater number of the more effective fac-
tors than in a Cream individual which happens to have just a
sufficient minimum of factors to insure the appearance of a tumor.
Therefore the intensity in the tendency towards the appearance
of tumors is greater in the English and 8| + 328 than in the
Cream. In all these individuals tumors appear, but the greater
intensity in the English and 8| + 328 causes the tumors to
appear earlier in life in these latter strains. There are, of course,
variations in the individuals of different strains, and some indi-
viduals of the 8J + 328 or English may resemble the average
Cream as far as the character and number of their tumor fac-
tors are concerned. There may be, in addition, among the

ORIGIN OF TUMORS IN MICE

221

multiple factors determining tumor growth, some special factors
which determine more directly the tumor age, and these special
factors may be inherited independently of the multiple factors
which determine mainly the appearance of a tumor.

These special factors may overcome the effect of the ordinary
tumor factors-, which in themselves already have a tendency to
affect the tumor age. If, on the other hand, we assume that
factors determining tumor age are always distinct from those
determining tumor incidence, then we must conclude that both
sets of factors are usually linked to each other in such a way that,
in general, in those strains in which the tumor rate is high the
factors determining an early appearance of tumors are also pres-
ent; but that it is possible to dissociate these two sets of factors
in certain cases, especially through hybridizations. Both in-
terpretations are related to each other; the first interpretation
including the second one. These interpretations must at present,
of course, be considered as merely of atentative character, but
they seem best to represent the facts so far established.

CONCLUSIONS

1. The tumor age of a certain strain is as definitely determined
by heredity as the tumor incidence ; the tumor age may be a finer
means of distinction between different strains than the tumor
incidence.

2. There is a definite relation between tumor rate and tumor
age in mice. In those groups or strains in which the tumor
incidence is great the tumors tend to appear early, and in those
groups or strains in which the tumor rate is low the tumors
tend to appear late. We can arrange our mice in three groups,
with high, medium, and low tumor rate; in these three groups
the tumor age shows corresponding changes.

3. The period of life at which a certain kind of tumor shows a
maximum frequency in a certain species is not definitely fixed,
but it varies with the rate of tumors in certain strains. The
usual statistics represent an average between the maxima in
different strains in which the maximum varies in accordance
with their tumor rate.

222

LEO LOEB

4. In addition to this general relation between tumor age and
tumor rate, there is in certain strains a specific tumor age which
may differ from that expected in the strain on the basis of its
tumor rate. Furthermore, in hybrid strains tumor rate and
tumor age may be inherited independently of each other.

5. These relations between tumor age and tumor rate can
best be explained if we assume that the hereditarily transmitted
constitution, so far as it represents the tendency of the organism
to develop tumors, depends on the cooperation of multiple fac-
tors. These multiple factors determine the intensity in the
tendency to tumor development in a certain individual. In
general, the greater is this intensity, the earlier do the tumors
appear and the greater is the probability that in related indi-
viduals there exists likewise a tendency to the development
of tumors.

It is furthermore probable that in addition to the general
factors determining the intensity in the tendency towards the
development of cancer, there exist factors which determine
specifically the tumor age in certain individuals and strains.

REFERENCES

(1) Lathrop, A. E. C, and Loeb, Leo: Proc. Soc. Exper. Biol, and Med., 1913,
xi, 34; Jour. Exper. M., 1915, xxii, 646 and 713, 1918, xxviii, 475;
Jour. Cancer Res., 1919, iv, 137.

FURTHER INVESTIGATIONS OF DISTURBANCES OF
BLOOD SUGAR EQUILIBRIUM IN THEIR
RELATION TO NEOPLASIA*

G. L. ROHDENBURG, O. F. KREHBIEL, and A. BERNHARD

From Columbia University, Institute of Cancer Research, founded by George Crocker f
F. C. Wood, Director, and the Laboratory of the Lenox Hill Hospital

Received for publication January 16, 1922

In previous publications (1), it has been shown that the
injection of a protein into the animal body is followed by dis-
turbance of the blood sugar equilibrium, and that this disturb-
ance of sugar equilibrium can be influenced (either increased
or decreased) by the ablation of one or more of the endocrines
(2). The present report concerns a further study of: (a) the
interrelationship of the endocrines in the regulation of blood
sugar equilibrium; (b) the relation of the degree and type of
blood sugar disturbance after the injection of protein to the
strength of the anti-bodies developed; (c) the relation of the
type and intensity of the reaction following the injection of
homologous protein to neoplasia in man.

I. EXPERIMENTAL PART

The technical steps of the method employed are simple.
The animal to be tested is starved for a period of twelve hours.
Just before, and again sixty minutes after the subcutaneous
injection of the antigen chosen, the blood sugar values are
determined in milligrams per 100 cc. The lower of the two
values is subtracted from the higher, the difference being taken
to indicate the degree of disturbance as measured in milligrams.
In the present work, in determining the blood sugar values

1 We beg to express our appreciation to all of the attending physicians and
surgeons of the Lenox Hill Hospital for the clinical material contained in this
report.

223

224 ROHDENBURG, KREHBIEL AND BERNHARD

we have used either the method of Epstein or of Kliner, or the
Wallace and Gallagher modification of the Folin-Wu method.
We have modified the Wallace and Gallagher method in that
the blood is measured instead of weighed, 0.1 cc. being used.
For this purpose a standardized pipet, so made that the capillary
bore necessary to contain the specified amount is about 15 cm.
long, was employed. It is essential that the same worker
make all determinations in a given experiment and that the
individual be skilled in the technical steps of the method chosen.
The same method for determining the sugar values should be
used throughout any one experiment. These conditions have
been complied with in the work recorded here.

In our early experiments it was found that in normal animals
after the injection of a given antigen, different types and degrees
of reaction occurred. For example, one animal might show an
increase of 100 mgm. in the blood sugar, while a second treated
exactly as was the first might show an increase of but 15 mgm.
A third animal, on the other hand, instead of showing an increase
might show a decrease, for example, of 25 mgm. Hereafter,
we shall use the terms "plus type" and " minus type" to desig-
nate, respectively, reactions in which the sugar concentration
rises, and reactions in which a decrease occurs.

In investigating these variations the first factor considered
was the influence of sex and age upon the type of reaction.
A number of young female, young male, old female, and old
male rats were tested in accordance with the method previously
outlined, the injected antigen being 0.5 cc. of a 1 per cent heterol-
ogous protein (beef peptone). The averaged results of this
experiment show that in the old males the plus and minus reactions
occurred with the same degree of frequency, while in young
and old females and in young males the plus reactions slightly
predominated.

In a second experiment, one or another of the endocrine
glands was removed, the test being performed at varying inter-
vals after ablation. A 1 per cent heterologous protein (beef
peptone), in doses of 0.5 cc, was first used as antigen, and after
an interval of several days a similar dose of a similar strength

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 225

of homologous protein (rat spleen) was employed. The glands
removed were as follows: both testes complete, both ovaries
complete, spleen complete, thymus complete, thyroid and em-
bedded parathyroids about 95 per cent entire; both adrenals
complete; pancreas about 90 per cent entire. As controls,
normal animals and animals from which one kidney had been
removed were used. There were from 6 to 12 animals in each
group.

The data of the experiments are given in table 1. The per-
centages of plus reactions are about the same in the normal
controls, in those with one kidney removed, and in the thymus-

TABLE 1

ORGAN REMOVED

PERCENTAGE GIVING PLU8
REACTIONS

PERCENTAGE GIVING MINC8
REACTIONS

Homologous
protein

Heterologous
protein

Homologous
protein

Heterologous
protein

Normal controls

Thyroid

100

Adrenals

100

Testes

100

free, pancreas-free, and ovary-free groups, irrespective of which
antigen was used. There was a difference in the types observed
in the spleen-free group, for of these 16 per cent gave a plus
reaction when injected with heterologous protein, as compared
with 83 per cent giving a plus reaction when homologous pro-
tein was used as antigen. One type of reaction only was ob-
served in three groups: the adrenal-free and thyroid-free groups
gave 100 per cent plus reactions, and the testes-free 100 per
cent minus reactions when injected with heterologous protein.
When homologous protein was used as antigen the adrenal-free
and testes-free groups reacted in the same manner, while the
thyroid-free groups gave 100 per cent minus instead of plus
reactions.

226

ROHDENBURG, KREHBIEL AND BERNHARD

Having observed the reactions after single gland removal,
and in that manner having obtained groups in which all the
animals reacted alike as regards type, we studied, in a third
experiment, the effect of temporary overactivity of various
glands. Extracts, in the proportion of 5 grams of the dried
powdered gland in 25 cc. of physiological saline solution, were
used as antigen, the dose being 0.5 cc. In this fashion extracts
of the pituitary, pancreas, adrenal, and thyroid glands of sheep
were prepared, and because these extracts, aside from their
supposed specific secretion content, were also heterologous
protein, a similar extract of dried sheep liver was used, as a
control. The other details of the experiment were similar to
those previously described. There were 12 non-operated rats
in each group. The data of this experiment are given in table 2.

TABLE 2

GLAND EXTRACT INJECTED

PERCENTAGE

PLUS REACTIONS

MINUS REACTIONS

Liver

Pituitarv

Pancreas , ,

Thyroid

In the animals injected with heterologous protein derived from
the adrenal and thyroid glands there was little or no change in
the percentages of the two reaction types as compared with the
group injected with heterologous liver protein. When pancreas
extract was injected, however, there was a predominance of the
minus reactions, which is in contrast to the pancreas-free ani-
mals in which injections of heterologous protein produced a
predominance of plus reactions.

The fact that, after single gland removal, but one type of
reaction occurred in certain groups led to a study of the rela-
tionship of the endocrine chain in its effect upon the reaction
types. Double gland ablations were done in such fashion that
all possible two gland combinations were obtained, always
excepting those with the pituitary.

BLOOD SUGAK EQUILIBRIUM AND NEOPLASIA 227

Six rats were used in each group. At varying periods after
operation these animals were injected with 0.5 cc. of homologous
protein and sugar estimations were made as before. The data
of the experiment are given in table 3.

When compared with single gland removal, double gland
ablation showed the following changes in type of reaction. The
spleen and testes combination showed results opposite to those
of spleen and ovaries, plus types predominating with the former
and minus types with the latter. The thymus-thyroid combina-
tion gave a predominance of plus types while the thymus-adrenal
gave a majority of minus reactions. Three groups showed one
type of reaction throughout; the adrenal-pancreas, testes-pan-

TABLE 3

TESTES

OVARIES

SPLEEN

THYMUS

PANCREAS

ADRENAL

Thyroid

100

Adrenal

100

Pancreas

100

Thymus

+ = plus reaction percentage
— = minus reaction percentage

creas, and the thyroid-ovary combination gave 100 per cent
minus reactions.

As has been stated in a previous paragraph not only did nor-
mal animals differ in type of reaction, but they also differed in
the intensity of the blood-sugar disturbances. If the intensity
of the reaction, as measured by the difference in milligrams per
100 cc. of the two sugar estimations, be compared in the young
and old (chart 1) it is found that the young react more distinctly
than do the old. Of the two sexes the males react more
distinctly than do the females.

If the intensity of the reaction which follows the injection
of heterologous protein be compared with that which follows
the injection of homologous protein in normal and in gland-free

228 ROHDENBURG, KREHBIEL AND BERNHARD

Oh a i;t 1

Comparison of Intensity of Disturbance of Dlood Sugar Jfyuilihrium
m Smgk Gf&nd'FreejJnimalSwilh Two Different Types ofAnhgen

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 229

animals (chart 2), it is found that the same degree of reaction
occurs with either antigen except in animals with the adrenals
or the thyroid removed. The adrenal-free group reacted more
distinctly to heterologous protein, while the thyroid-free reacted
more intensely to homologous protein. Spleen-free animals
showed a lessened intensity of reaction to both antigens.

When the reaction following injection of homologous protein
is studied in animals from which two glands have been removed,
the results (2), to quote from the paper previously referred to,
are as follows: " There are two combinations of endocrines
which when ablated inhibit the reaction; these are (1) spleen
and either gonad set, and (2) adrenal and pancreas. If one
gland from either system be ablated and with it one gland from
the other system, there results little or no change from the nor-
mal reaction. If, however, one gland from one system be
removed and with it one gland from another system, and in
addition either the thyroid or thymus, then marked inhibition
occurs. Removal of one gland from one system and either
thyroid or thymus does not produce inhibition. If both glands
be removed from each system, then, in order to induce inhibi-
tion, both of the intermediate glands, i.e., thyroid and thymus,
must also be removed; the ablation of one intermediate gland
is not sufficient. From this it appears that there are two sys-
tems which control the reaction and these systems are connected
through two apparently interchangeable glands.''

Stimulation of the reaction with double gland ablations oc-
curred in two systems, analogous to those producing inhibition.
These systems were the thyroid in combination with either
ovaries, pancreas, or spleen, and the pancreas in combination with
either testes, thyroid, or spleen.

The exact significance of variations in type and in intensity
of reaction is somewhat in doubt. It might be supposed, since
disturbances of blood-sugar equilibrium follow the injection of
an antigen until the injected animal has exhausted the power
to produce further antibodies against the antigen injected,
that variations in the type or in the intensity might be indica-
tive of the relative strength of the antibodies formed.

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 3

230 ROHDENBURG, KREHBIEL AND BERNHARD

Whether such an hypothesis is valid or not was determined in
the following experiment. Rabbits, in groups of four animals
each, were prepared as follows: one group served as normal
controls; from a second the testes were removed; while from a
third the complete thyroid apparatus was excised. One week
after operation and after twelve hours' starvation blood-sugar
estimations were made, the animal was then injected intraperi-
toneally with 2 cc. of a 50 per cent suspension of washed sheep
red cells, and one hour after the injection the blood sugar was
again determined. Lytic titers were made with serum obtained
from the first bleeding. Two subsequent injections and double
sugar determinations as well as lysin titers were made at weekly
intervals, and a final lytic titer was made one week after the
last injection. As is shown in chart 3 the thyroid-free animals
showed a greater sugar disturbance than did the control animals,
and the control animals in time showed a greater disturbance than
did the testes-free animals. If intensity of reaction bears relation
to the strength of antibody formation then we should expect
the thyroid-free group to show the highest lytic titer, and the
testes-free the lowest titer. Such, however, was not the case,
for the normal controls gave the highest titer, the testes-free
next, and the thyroid-free the lowest titer.

There being no demonstrable relationship between the intensity
of the reaction and the strength of the antibody development,
the most plausible remaining explanation is that the phenom-
enon is an indication of the relative activity of various of the
endocrines. The failure of the injected animal to react after
the maximum antibody production has been attained might also
be taken as an indication of the temporary exhaustion of the
endocrines.

It is more logical, in endocrine studies, to consider averaged
group behavior rather than individual behavior, and it is on
averaged group behavior that the scheme of interrelationship
given below has been worked out. If the ablation of a gland
or of a group of glands is followed by a minus reaction, then
that gland or the combination when in situ may be considered
as causing a plus reaction. Similarly, if a given gland or gland

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 231

group removal brings about an inhibition, then that particular
combination when in situ could be considered as stimulators
and vice versa. The deviation one way or another from the
behavior of the control group of normals would serve in lesser
degree to indicate further interglandular relationships.

Comparison oi Lyixc Tj4st a.22 d In iensziy 0/

Thus as is diagrammatically shown in chart 4, if the injection
of a homologous protein is followed by a drop in the blood-sugar
concentration, either the thyroid or, if the individual be a male,
the testes, may be considered as inefficient; whereas, if the sugar
values rise, the adrenals may be considered as overactive. It

232

ROHDENBURG, KREHBIEL AND BERNHARD

may also be supposed that decreases in the sugar concentration
are due to lessened activity of the adrenal whose function is
to keep up the sugar concentration. This insufficiency may be
due either primarily to the adrenal or secondarily to the inef-
fectiveness of the stimuli reaching that gland through the spleen,
thymus, or pancreas. In a similar manner, a rise in the blood-
sugar concentration may be attributed to failure of the thyroid
or testes to hold the blood sugar at the lower level, and this

may in turn be due to the thyroid or testes primarily, or it may
occur secondarily because of failure of the stimuli from either
spleen, pancreas, or thymus.

The intensity of the reaction has been worked out on similar
lines. Accepting, for example, that a normal intensity of reac-
tion for the rat is 31 mgm., if a gland removal results in reac-
tion intensity of but 10 mgm. then the gland removed may be
considered when in situ to be one of those helping to keep up

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 233

the normal intensity. If two glands be removed from two dif-
ferent groups and it is found that group A has a reaction intensity
of 45 mgm. and group B a reaction intensity of 10 mgm. it may
be accepted that the gland whose removal brought about the
increased intensity, is when in situ one of the group which keeps
the intensity down, the opposite being the case in the group in
which gland removal caused a less intense reaction. Now, if
both glands be removed from a third group and the reaction
intensity of this group is found to be 10 mgm., it may be supposed,
gland A in situ being an inhibitor, and gland B in situ being a
stimulator, that gland A sends an inhibiting influence to gland
B, for gland A being removed permits gland B to act without
restraint, and the reaction is intensified. Gland B being re-
moved permits gland A or some other gland to act without
restraint in consequence of which the reaction intensity falls.

The intensity of reaction according to glandular interrela-
tionship is indicated in chart 5. For example, if the adrenal
be removed, the spleen lacks a stimulating influence and the
thyroid an inhibiting one; the retention of a normal reaction
intensity might, therefore, be expected, for the plus and minus
influences balance each other, the actual experimental values
being: normal group, reaction intensity 31 mgm.; adrenal-free
group, reaction intensity 36 mgm. Ablation of the spleen re-
sults in the removal of secondary stimulation received via the
gonads, adrenal, and pancreas, and the removal of a stimulating
impulse sent to the pancreas, as well as directly upon the blood-
sugar concentration. It might be expected, therefore, that the
splenic removal would depress the intensity of the reaction.
The actual experimental values obtained were: normal animals,
reaction intensity 31 mgm.; spleen-free group, reaction intensity
10 mgm.

Theoretically, therefore, if the blood sugar concentration
after an injection of homologous protein falls distinctly, it
may be supposed that the mechanism which causes increased
concentration (chart 4) has not responded or that the mechan-
ism which depresses has been stimulated. An intense reaction
would speak for thyroid inhibition. In contrast, if the reaction

234

ROHDENBURC, KREHBIEL AND BERNHARD

were slight, say 10 mgm., splenic inhibition would be suspected.
Inversely, if the sugar concentration rises distinctly it would
indicate adrenal hyperactivity in combination with thyroid
and pancreas, while if the reaction were minimal it would indi-
cate adrenal hyperactivity in combination with spleen and
gonad.

That disturbances of blood-sugar equilibrium no longer occur
after the injection of protein, when the organism has attained

its maximum antibody producing power is, on this basis, to be
interpreted as indicating temporary exhaustion of certain of the
endocrines.

CLINICAL APPLICATION

It has already been noted that a majority of the mice bearing
spontaneous tumors subjected to this test have shown a de-
creased intensity of reaction. In order to follow this lead,
further investigations were carried out on human beings suf-

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 235

fering from a large variety of pathological conditions. The
technic of the test as applied to the human cases was as follows:

After the evening meal no medication was given, the patient
was starved until the last specimen of blood was withdrawn the
following morning. Just before and again 45 and 120 minutes
after the subcutaneous injection of 10 c.c. of homologous pro-
tein, blood was withdrawn and the sugar content determined.
The degree of disturbance of equilibrium was estimated by
subtracting the lowest from the highest of the three values
obtained, the difference in milligrams being the standard of
comparison. The homologous protein used was either blood
serum or ascitic fluid obtained from individuals free from
syphilis, tuberculosis, or cancer, in order to obviate any possible
chance of disease transmission. The sterility of the fluid was
assured by culture. No reactions were encountered as a result
of such injections except in one instance where there was a slight
chill followed by a rise in temperature to 102°F. With the
Folin-Wu method we have found that not more than three
unknowns can be compared with the same standard because
the colors of the standard fade rapidly.

The clinical diagnoses have in every instance been confirmed
either by histological or bacteriological examination, autopsy,
operation, or investigation of the subsequent course of the
patient. All the cases of neoplasia have been confirmed
microscopically.

Before proceeding to an analysis of the data obtained as a
result of these investigations it is profitable to revert to those
obtained during investigations already reported (3). In this
previous investigation blood-sugar equilibrium was disturbed
by the oral administration of 100 grams of anhydrous glucose.
When the individuals examined by this method were divided into
male and female, cancerous and noncancerous, the following
results were obtained. There was no predominance of type in
one group over another, i.e., a particular type of reaction could
not be said to be characteristic of any given group. If the
intensity of the reaction, i.e., the degree to which blood-sugar
equilibrium was disturbed was compared (chart 6) it was found

236

ROHDEXBURG, KREHBIEL AND BERXHjLRD

that with the exception of the non-cancerous males all the groups
gave the greatest percentage of cases with a reaction intensity
of from 61 to 120 mgm.; in brief, there was nothing characteristic
of neoplasia.*

However, when the data of the types and intensities of reac-
tion observed after the injection of homologous protein in man
were similarly analyzed there were some very interesting and
suggestive variations.

Comparison of Type and Jntensiiy of disturbance of Blood Sugar
after Glucose Ingestion in Cancer and Non-Cancer Bearers.
Mgrrrs o ZD 40 SO N 100 m ufo /60 OO zoo

X
zo

Qfsz non-cancer cases ss% gave plus reactions.

Of Z4 cancer cases 3S%gave joins reactions.

Intensity of Jte action.

Cancer Cases

Nonm C&ncer Cases.

Chart 6

Table 4 gives the statistical data of the 209 cases which have
been examined. When this information is presented as in
table 5, several points are evident; a larger percentage of the

* In a recent publication Killian and Kast (4) stated that the hyperglycemia
observed in cases of neoplasia is attributable to the nitrogen retention present
in many of these cases. This factor had not been considered by us in our previous
communications. A review of our data, however, does not bear out the conten-
tion advanced by these authors, but confirms in a very emphatic manner the
observation made by Spence (5) who attributed the hyperglycemia to the age of
the individuals. In his investigations Spence showed that hyperglycemia occurs
with increasing frequency in the more advanced age periods.

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 237

TABLE 4

DIAGNOSIS

General carcinomatosis

Carcinoma rectum

Carcinoma lung

Carcinoma stomach —
Multiple myelomata —
Carcinoma stomach —

Carcinoma cecum

Carcinoma pancreas —

Carcinoma rectum

Carcinoma gall bladder

Carcinoma breast

Melanocarcinoma skin.

Carcinoma stomach

Sarcoma lung

Carcinoma stomach

General carcinoma

Sarcoma femur

Carcinoma colon

Carcinoma stomach . . .

Carcinoma rectum

Carcinoma general

Carcinoma stomach

Carcinoma general

Carcinoma stomach

Carcinoma rectum

Carcinoma stomach

Carcinoma cervix

Carcinoma stomach

Carcinoma general ,

Carcinoma stomach

Carcinoma sigmoid

Carcinoma bladder

Carcinoma stomach

Sarcoma ovary

Carcinoma vagina

Carcinoma bladder

Carcinoma prostate

238

ROHDENBURG, KREHBIEL AND BERNHARD

TABLE 4— Continued

Carcinoma stomach

Carcinoma liver

Carcinoma stomach

Carcinoma esophagus...

Carcinoma rectum

Carcinoma breast

Carcinoma generalized..
Carcinoma generalized..

Carcinoma cervix

Carcinoma breast

Carcinoma liver

Carcinoma esophagus...

Carcinoma colon

Carcinoma esophagus. . .

Sarcoma parotid

Sarcoma pelvis

Sarcoma retroperitoneal

Carcinoma stomach

Carcinoma breast

Carcinoma rectum

Carcinoma stomach

Epithelioma cheek

Carcinoma stomach

Carcinoma larynx

Carcinoma stomach

Carcinoma cervix

Carcinoma intestine

Carcinoma liver

Carcinoma cervix

Carcinoma tongue

Carcinoma breast

Carcinoma colon

Carcinoma breast

Carcinoma sigmoid

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 239

TABLE i— Continued

DIAGNOSIS

Carcinoma stomach

Carcinoma breast

Carcinoma stomach

Carcinoma esophagus —
Retroperitoneal sarcoma
Carcinoma esophagus —

Carcinoma stomach

Carcinoma lung

Carcinoma esophagus

Sarcoma tendon sheath..

Carcinoma colon

Carcinoma stomach

Epithelioma penis

Carcinoma uterus

Lymphosarcoma neck

General carcinoma

Carcinoma ovary

Carcinoma bladder

Carcinoma mouth

Carcinoma general

Carcinoma appendix

Carcinoma uterus

Carcinoma testes

Carcinoma stomach

Carcinoma lip

Endothelioma brain

Syphilis

Inguinal hernia

Encephalitis

Biliary cirrhosis

Encephalitis

Syphilis

Neuritis

Tuberculosis

Achylia gastrica

Gastric ulcer

Orchitis

Varicose veins

Leukemia

Fibroadenoma breast

Cryptorchidism

DISTURBANCE

BLOOD SUGAR

mgm.

-3

-9

-33

-35

-51

+54

+20

+36

-34

-42

-27

+30

+28

-34

-33

-31

+30

-21

+19

+22

-20

+49

+14

+30

-48

-44

-24

-15

-16

+21

+42

+15

-21

+15

+18

-42

+18

-39

240

ROHDENBURG, KREHBIEL AND BERNHARD

TABLE 4— Continued

Cryptorchidism

Angioneurotic edema...

Hy perthy roid ism

Retinitis pigmentosa. . .

Pernicious anemia

Orchitis

Ovarian cyst

Endocarditis

Myocarditis

Cardio-nephritis

Lymphadenitis

Fibromyoma uteri

Endocarditis

Gastritis

Postoperative adhesions

Pregnancy

Hypertrophied prostate

Fistula in ano

Catarrhal jaundice

Graves' disease

Pernicious anemia

Subacute gastritis

Catarrhal jaundice

Endometritis

Oophoritis

Neurasthenia

Gastric ulcer

Splenomegaly

Cholelithiasis

Mixed tumor parotid

Cerebral thrombosis

Diabetes

Uterine fibroids

Papilloma bladder

Cholelithiasis

Diverticulitis

Gastric ulcer

Neurasthenia ,

Tuberculosis

Ovarian cyst

Cholecystitis

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA

241

TABLE 4— Concluded

DISTCKBANCE
OF

BLOOO SCGAH

Pyelitis

Cholecystitis

Sciatica

Chronic inflammation

Gastric ulcer

Tuberculosis

Chronic cholecystitis

Gastric neurosis

Cholecystitis

Angina pectoris

Gastric ulcer

Lung abscess

Pernicious anemia

Pregnancy

Gastric ulcer

Cystocele

Duodenal ulcer

Orchitis

Syphilis

Cholecystitis

Duodenal ulcer

Pernicious anemia

Chronic appendicitis

Nephritis

Bronchopneumonia

Fibroid uterus

Fissure in ano

Prolapse uterus

Diverticulitis

Duodenal ulcer

Fibroid uterus

Nephritis

Cholecystitis

Chronic endocerricitis. . .

Cholecystitis

Tuberculosis

Hepatitis

Cholelithiasis

Polyp of sigmoid

Hypertrophy of prostate

Septic endocarditis

Pernicious anemia

242

ROHDENBURG, KREHBIEL AND BERNHARD

cases of neoplasia of the gastro-intestinal tract and of the liver
and gall-bladder show reactions of an intensity of less than 12
mgm., than is the case in individuals who have non-malignant
diseases of these organs. In diseases of the genito-urinary
apparatus this relationship does not apparently hold true, but
in the series of miscellaneous neoplasias, no one type being
present in sufficient numbers to warrant separate tabulation
it is again evident that a larger proportion of the cancerous
individuals show a lessened intensity of reaction than do a
similar group of non-cancerous individuals.

TABLE 5

REACTION

Less than

More than

12 mgm.

7 (54%)

6 (46%)

Neoplasia of gastrointestinal tract

50 (88%)

7 (12%)

Non-grouped neoplasia

29 (74%)

10 (26%)

Non-malignant disease of genitourinary tract. .. .

9 (41%)

13 (59%)

Non-malignant disease of gastrointestinal tract . .

13 (38%)

21 (62%)

Non-grouped non-neoplastic diseases

9 (20%)

35 (80%)

All cases of neoplasia

86 (78%)

23 (22%)

All cases non-neoplastic

31 (31%)

69 (69%)

These differences are distinct when compared with the simi-
larity of behavior found in our previous series where blood-sugar
equilibrium was disturbed by the ingestion of glucose. The
observation is possibly of value as a diagnostic procedure. In
diseases involving the gastro-intestinal tract, from mouth to
anus, duodenal ulcer excepted, and in diseases involving the
gall-bladder and liver, 88 per cent of all the cases of neoplasia,
i.e., 50 of 57 cases, showed a reaction of less than 12 mgm. As
regards the reaction types it was found that 39 per cent of the
neoplasia cases gave plus reactions as compared with 52 per cent
in the non-cancerous group. If the previously outlined endo-
crine relationship be accepted, such lessened reaction would
indicate either spleen, adrenal, or gonad insufficiency, or thyroid,
adrenal, and pancreas overactivity.

BLOOD SUGAR EQUILIBRIUM AND NEOPLASIA 243

CONCLUSIONS

1. The disturbance of blood sugar equilibrium which follows
the injection of a protein varies in intensity and in type.

2. The removal of certain of the endocrines influences both
the intensity of the reaction and its type.

3. Neither the intensity of the reaction nor the type have
any relation to the strength of the antibody development after
the injection of a selected protein.

4. A study of the type and reaction intensity has been used
in an attempt to establish paths of endocrine correlation.

5. Human cases of neoplasia show a larger proportion of
weak sugar reactions, i.e., of 12 mgm. or less, than do non-
cancerous individuals. This is particularly true of neoplasms
of the gastro-intestinal tract and liver.

REFERENCES

(1) Rohdenburg and Pohlman i Am. Jour. Med. Sci., 1920, clix, 853.

(2) Rohdenburg and Krehbiel: Am. Jour. Med. Sci., 1921, clxii, 28.

(3) Rohdenburg, Bernhard, and Krehbiel: Am. Jour. Med. Sci., 1920.

clix, 577.

(4) Killian and Kast: Arch. Int. Med., 1921, xxviii, 813.

(5) Spence: Quart. Jour. Med., 1921, xiv, 314.

PREGNANCY AND TUMOR GROWTH

ISIDOR KROSS

From Columbia University, Institute of Cancer Research, F. C. Wood, Director
Received for publication January 16, 1922

The relationship between pregnancy and tumor development
has been investigated by a number of observers, whose results
have been, however, in many instances contradictory. For
example, Moreau (1) and Herzog (2) found that transplanted
tumors grow with unusual rapidity in pregnant animals. Haaland
(3), on the other hand, maintained that pregnancy exerts a
markedly restraining influence on the growth of transplanted
tumors, which continue to develop in their usual fashion after
parturition. Bashford and Murray (4) hold that " pregnancy' 1
in the female "and full sexual activity in the male constitute no
bar to successful transplant ation." Uhlenhuth and Weidanz
(5) observed retardation in the growth of tumors during
pregnancy and also more frequent spontaneous retrogression.
Bridre (6) noticed a low percentage of positive inoculations in
pregnant animals, and Ehrlich (7) observed that inoculations
into animals bearing young was frequently followed by negative
results and that in those successfully inoculated the growth of
the tumor was greatly retarded. Then, again, Albrecht and
Hecht (8) maintained that pregnancy affects the growth of
tumors just as little as the presence of a tumor influences con-
ception or gravidity. Cuenot and Mercier (9) reported a most
interesting observation. They noticed that BorreFs tumor "B,"
which rarely underwent spontaneous absorption, if inoculated
before fecundation developed during gestation and receded dur-
ing lactation. The tumor, however, did not regress if one mouse
only was born, thus leaving the activity of the mammary gland at
a minimum. Neither did absorption occur even in the presence
of several young, if the tumor was so situated that its vasculari-

245

246

ISIDOR KROSS

zation was independent of that of the mammary gland. Fichera
(10) explained these various inconsistencies by assuming that
when many embryos were present the food stuff was almost wholly
consumed by them, while if there were only a few some of this
material was available for the tumor cells.

Slye has concluded from her observations of spontaneous
tumor growth in the mouse that: First, reproducing females
grow much less tumor than do non-reproducing females of the
same age, etc. Second, reproducing females grow much less
tumor while they are reproductive than they do while they are
non-reproductive. In other words, she finds that pregnancy
exerts a markedly retarding influence upon growth of tumor.

In re-investigating this problem rats were employed. The ani-
mals selected were young adults, sexually mature, vigorous, and in
very good condition, which had been separated from male animals
for a period of one month to avoid the possibility of unrecognized
pregnancies. The rats were mated with strong, vigorous young
males that were kept in the same boxes not only during the entire
period of gestation but also after the females had littered, in
order to impregnate the same animals immediately after par-
turition. In all cases the date of conception was calculated by
counting back twenty-one days before the date of littering. All
the animals were inoculated on the same day with 0.003 gram of
Flexner's rat carcinoma, the time chosen being two days after
the animals had been mated. In all, 105 animals were inoculated.
Of these, 43 (41 per cent) became pregnant and littered normally.

In analyzing the cases in which no growth occurred, it is interest-
ing to note that about 7.8 per cent of the males inoculated showed
no growth, and that about 8.4 per cent of the females showed no
growth. In the group of females without tumor growth, 44 per
cent did not become gravid. The other 46 per cent did. These
figures tend to show that in the negative cases neither sex nor
gravidity played any role.

Tracings of the tumors were made every four days, and in
this way growth was depicted simply but effectively.

A comparison of the charts (figs. 1 to 4) of the tumors of the
pregnant and of the non-pregnant groups shows at a glance

PREGNANCY AND TUMOR GROWTH

247

that in these experiments pregnancy had absolutely no effect at
all upon the growth of the transplantable Flexner rat carcinoma.

•

10 CM.

Fig. 1

CONTROL.

• •000

•-•■•••.©
•••••#

• • • • » ©

•••••••

• • • • -t

• • • • •
• • » •

• • • • • 4

• ••00

• • • t f

• * \ \ X %

to cm.

Fig. 2

DISCUSSION

It seems impossible completely to reconcile all of the conflicting
results which have been published in regard to the effect of
pregnancy on the growth rate of inoculated tumors, chiefly be-
cause we have insufficient data upon which to base any conclusions

248 ISIDOR KROSS

as to the growth capacity of the tumors which the various
experimenters have employed. Nor do we know, in many in-
stances, the site of inoculation. Mice are especially unsuitable
animals for such a study, as the implanted tumors are apt to be

FREaHAHT control.

- — — • •

- -

— •

ft ft

— — — • •

•

ft ®

_ _ _ . •

•

• •

* ft

•

• •

•

ft ft

*
X

•

• 9

•

« f

•

• •

• • •

•

♦

ft •

•

• •

0 •

•

—

• ft

• •

•

— - - • •

•

• ft

•

• •

f f

•

# *

x - LtTTEKtP

• i

•

'0 CM. |0 CM.

Fig. 3 Fig. 4

variable in their growth rate, and may become so large as to kill
the animal quickly. This obviously complicates the problem
and renders careful control and statistical analysis of the results
necessary. But it does seem possible to draw a few conclusions,

PREGNANCY AND TUMOR GROWTH

249

not entirely inharmonious with the reported experimental work
and the facts obtained by clinical observation.

The tumor grafts in the experiments here reported were not
placed in the mammary gland but by intent in the axilla; they
were, therefore, not affected by the increase in the vascularity of
the breast at the beginning of pregnancy as they would be if
placed directly in the main mass of the mamma. This increase
in vascularity is undoubtedly the explanation of the rapid growth
of tumors in the breast observed in human beings. But the
conditions in the human female and in female mice are not
comparable. A mouse weighing 15 to 20 grams will frequently
carry a tumor weighing one-tenth of its weight. Unfortunately,
the weights of the tumor-bearing animals are not given in Miss
Slye's paper, though the dimensions of the tumors are furnished.

On the other hand, the weight of the tumor in a human female
is rarely more than one one-hundredth or one one-hundred and
fiftieth of the body weight. Mammary cancer in a mouse, there-
fore, may correspond in mass to a rapidly growing breast car-
cinoma of 10 to 15 pounds in a human female, a tumor which is
practically never seen. It is quite possible that the demand for
food for such a theoretical tumor if complicated by pregnancy
might result in the starvation of the tumor for a time as Moreschi
(12) has shown to be possible.

It is, therefore, evident that Miss Slye's results, while applica-
ble to one species of mammalia under the conditions which she
specifies, are by no means obviously transferable even to other
types of rodents, much less to human beings.

The fact that in the additional experiments here reported no
influence on the growth rate of tumors was shown is probably
due to the disproportion between the size of the rat and its em-
bryos, frequently weighing 150 grams with 25 grams of embryos,
and the tumors, which are usually not more than one-fiftieth
to one-hundredth of the weight of the animal. So, too, in
mice the proportionate weight of the embryos demanding food
is larger than that ordinarily occurring in the human female,
a child rarely being over one-fifteenth of the mother's weight,
while in the multiple mouse pregnancies 5 grams of embryos may
be produced at a single birth from a mother weighing only 22
grams.

250

ISIDOR KROSS

A grafted tumor, such as we have employed, being a mass of
somatic cells growing in a foreign host, affords a possibility for
studying the problem of tissue growth independently of organ or
other relationships. The tumor is simply implanted in the sub-
cutaneous connective tissue, while a primary tumor of the breast,
in its inception at least, has important anatomical correlations
with the tissue in which it grows. It is thinkable that some of
the results observed by Miss Slye may be due to changes in the
connective tissue stroma of the tumor rather than to the epithe-
lium, but in the implanted tumor the connective tissue is furnished
by the host, the epithelium being derived from another animal,
so that the situation is quite different.

It seems safe to conclude, therefore, from our own experi-
ments, that pregnancy of itself does not necessarily alter tumor
growth rates, but that such interference, when it is observed, is
due to the forced division of food substance between the tumor
and the offspring; similar checking of the tumor can be observed
in implanted tumors when the food of the non-pregnant animal
is reduced approximately to the starvation point.

A comparable diminution in the tumor growth rate may be
seen in human beings in the terminal stages of extreme cachexia
with innutrition but, as a rule, in man pregnancy either has no
effect on the progress of a cancer or, if the mammary or uterine
tissues are involved, hastens the growth.

There is nothing, therefore, in the varying results of published
experimental work which cannot be harmonized or which con-
troverts the clinical observations already recorded.

REFERENCES

(1) Moreau: Arch, de mod. exp6r. et d'anat. path., 1894, vi, 693.

(2) Herzog: Jour. Med. Res., 1902, n.s. iii, 76.

(3) Haaland: Berl. klin. Wchnschr., 1907, xxiii, 718.

(4) Bashford and Murray: First Sci. Rep., Imperial Cancer Research Fund,

London, 1904, p. 14.

(5) Uhlenhuth and Weidanz: Arb. a. d. k. Gsndhtsamte., 1909, xxx, 440.

(6) Bridre: Ann. de l'lnst. Pasteur, 1907, xxxi, 763.

(7) Ehrlich: Verhandl. d. deutsch. path. Gesellsch., 1908, xii, 29.

(8) Albrecht and Hecht: Wien. klin. Wchnschr., 1909, xxii, 1783.

(9) Cuenot and Mercier: Compt. rend. Acad. d. sc., 1909, cxiix, 1012.

(10) Fichera: Jour. Exper. M., 1911, xiv, 320.

(11) Slye: Jour. Cancer Res., 1920, v, 25.

(12) MoREseni: Ztschr. f. Immunitatsforsch., Orig., 1909, ii, 651.

IS CANCER MORTALITY INCREASING?

WENDELL M. STRONG
Received for publication March 17, 1922

To determine whether or not there is a real increase in cancer
mortality we must compare cancer death rates for the same ages,
since cancer is an old age disease. The effect of age distribution
is so great that it is safe to say that any considerable increase in
the proportion of the population at the older ages, for instance,
the emigration of the young men and women, would cause a
noticeable increase in the cancer deaths per 100,000 population,
while an influx of young men and women would decrease them.

Dr. Frederick L. Hoffman, statistician of the Prudential Life
Insurance Company, in a paper on "The Menace of Cancer"
published in 1913, and in his comprehensive book on "The Mor-
tality from Cancer" published in 1915, has given the statistics
available at the time of writing in a very complete form. These
statistics were drawn both from the United States and abroad.
From these I have selected a few.

In the registration of 1900, the cancer death rate per 100,000
population, which will hereafter be spoken of as the cancer death
rate or cancer mortality, increased from 63 in 1900 to 79 in 1913,
an increase of over 25 per cent. England and Wales for the same
period showed an increase from 83 to 105 and most other Euro-
pean states showed a very considerable increase. Going back
forty years, the rate in twenty large American cities was 49 for
the period 1881-1885; for 1913 it had increased to 89. In twelve
European countries the rate was 69 for the period 1896-1900;
for the period 1906-1910, it had increased to 81. These were
general population statistics without subdivision by ages. There
were also available, however, some extensive statistics subdivided
by ages. The cancer death rate in the registration states of the
United States of 1900 subdivided into ten-year age groups, 25-34,

251

252

WENDELL M. STRONG

35-44, etc., for 1901 compared with 1911 shows an increase in
the rate for each age group for the later year, such increase being
very considerable for ages above 45, which are the significant
ages in cancer. Massachusetts, for the two periods 1901-1905
and 1906-1910, with a different distribution into age groups and
subdivided by sexes, showed an increase for each adult group and
for each sex separately, in the later period the increase again
being very considerable for the older groups. While these are
but a few of the statistics, they are illustrative of the trend of
practically all, and, standing alone, would appear to indicate a
great and rapid increase in cancer mortality. We have, however,
later and very extensive statistics from two of the great insurance
companies which are not in accord with the above.

In 1919, Dr. Louis I. Dublin, statistician of the Metropolitan
Life Insurance Company, published in this Journal (1) the
experience of that company on its millions of industrial policy
holders for the years 1911-1916. Tables 1 and 2 are from this
experience. They show fluctuations but no definite trend.

Table 1 is for the age group 55-64 only, which group Dr. Dublin
chose as illustrative. We find in it the lowest mortality for white
males in 1912 and 1915, the highest in 1913 and 1914; for white
females the lowest in 1911 and 1914, the highest in 1912 and 1915.
It is a curious accident that the years of low mortality for the one
sex are in three cases the high for the other.

Table 2 compares the first two years of the period with the last
two and in total shows but one per cent difference between them,
a difference smaller than would be expected from merely acciden-
tal fluctuations. The two features which may be of significance
are that the total of males shows an increase of 5 per cent and
that the differences in each age group, as well as in the total of
all ages (except perhaps 65-69), are no more than would be ex-
pected from accidental fluctuations.

The cancer experience for the Mutual Life Insurance Company
for the years 1915-1920 has just been complied by Dr. Brandreth
Symonds, chief medical director. This experience, given in
table 3 below, is not subdivided by sexes, but as the great pre-
ponderance of insurance was on white male lives it may be taken
as representing white male cancer mortality.

IS CANCER MORTALITY INCREASING?

253

TABLE 1

Metropolitan Life Industrial Department. Cancer death rate per 100,000,

ages 55 to 64-

YEAR

ALL CLASSES

"WHITE

COLORED

Males

Females

Males

Females

1916

386.4

358.0

427.4

218.3

339.9

1915

380.8

336.0

427.8

175.7

394.3

1914

390.9

385.0

423.3

167.7

351.7

1913

384.1

370.3

414.6

195.2

368.3

1912

381.9

334.1

443.2

176.4

325.4

1911

368.7

353.3

400.2

158.0

373.7

TABLE 2

Metropolitan Life Industrial Department. Ratio of cancer death rate for years
1915-1916 to death rate for years 1911-1912 by percentages

ACE PERIOD

ALL CLASSES

WHITE

COLORED

Males

Females

Males

Females

25 and over

per cent

101.0

per cent

105.2

per cent

98.3

per cent

105.1

per cent

100.7

25 to 34

98.0

121.0

94.1

95.5

95.0

35 to 44

100.1

104.3

96.9

142.6

98.2

45 to 54

99.1

109.7

97.0

67.7

93.1

55 to 64

102.2

101.1

101.2

118.0

105.2

65 to 74

107.0

115.9

101 .9

117.2

107.0

75 and over

101.0

104.1

104.4

110.7

48.6

TABLE 3

Mutual Life Insurance Company. Cancer death rate per 100,000

AGE PERIOD

1915

1916

1917

1918

1919

1920

ALL YEARS

1915-1920

25-29

15.34

3.98

0.90

7.53

1.47

5.08

5.35

30-34

11.52

15.19

19.55

10.54

5.52

5.49

10 .55

35-39

7.61

12 .95

13 .38

10.99

5.74

18.69

11.87

40-44

31.39

45.31

30.86

26.92

33.73

21.21

31.07

45-49

40.36

41 .82

51 .59

60 96

51.56

78.80

55.15

50-54

165 .47

157.35

100.67

79.25

73.45

70.02

105 .78

55-59

132 .22

213.84

204.03

153.07

152 .73

159 .95

168 .88

60-64

235 .74

366 .73

255 .24

308.53

285 .56

272 .73

287 .21

65-69

225 .77

380.94

532.84

357 .91

502.42

423 .88

407 .99

70 and over

554.09

1000.67

834 .83

856 .57

731 .75

982 .20

829.54

254

WENDELL M. STRONG

If the first two years, 1915-1916, combined are compared with
the last two, 1919-1920, we find that up to age 65 the cancer
mortality ran in general considerably lower for the later than for
the earlier period, while for ages 65 and older the reverse was
true. Of the nine five-year groups (including "70 and over" as
such a group) , beginning with age 30-34, five show a higher mor-
tality in the earlier two-year period and four in the later. Had
the subdivision into age groups been the ten-year groups of the
Metropolitan, beginning with 25-34, every group up to age 65
would evidently have shown a higher mortality in the earlier
period.

The year 1915, for some unexplained reason, was a year
of very low cancer mortality in the Mutual, a peculiarity
which did not appear in the Metropolitan experience and
which was, therefore, probably accidental. It is consequently
worth while to see what would have been the result had 1915
been eliminated so that we should have compared the years
1916-1917 with 1919-1920. We find that of the nine five-year
groups, seven would have shown a higher mortality for the earlier
period and only two for the later; moreover, had the grouping
been by ten-year groups, as in the Metropolitan experience, each
ten-3rear group from age 25 up would evidently have shown a
higher mortality in the earlier period than in the later.

Again, a comparison of the last two years combined with the
preceding four years combined, shows that of the nine groups
five had a higher mortality in the earlier period; we must recog-
nize, however, that two of the four were 65-69 and "70 and over"
which are of greater importance than groups near the lower age
limit. Had the exceptional year 1915 been eliminated, however,
the result would have again looked extremely favorable for the
later period.

In order to have the most recent possible data, the approxi-
mate cancer mortality for the year 1921 has been worked out,
although exact figures are not yet available. The 1921 results
compared with the average for the preceding six years show a
lower mortality for each of the seven five-year age groups from
30-34 to 60-64, but a higher mortality for 65-69 and "70 and
over."

IS CANCER MORTALITY INCREASING?

255

Emphasis is given to the failure of the Mutual experience to
show any tendency to an increasing death rate by the fact that
the lives involved were nearly all white males, and in other ex-
periences where the sexes are separated any tendency towards
increase has generally been more among males than females.

The Metropolitan experience has been extended from 1915
through 1920 but has not yet been made public. Through the
courtesy of Dr. Dublin I have had an opportunity to see these
statistics and to include in this paper the deductions from them.
They show in general the same lack of any tendency to increase
in the age groups from age 30 to age 65, but they show an increas-
ing tendency from age 65 upwards.

Thus, we have had the experience of the two companies cover-
ing the eleven-year period from 1911 to 1921, inclusive. The data
in these experiences should probably be at least as accurate as
to the causes of death as those of any other experience, because
the insurance companies try to obtain accurate knowledge of
the cause of death at the time the claim is paid. The figures
indicate, taken at their face value without considering whether
there are any modifying influences, for ages below 65 either a
fluctuating or a slightly decreasing cancer mortality; for ages 65
and over they vary, but probably, on the whole, indicate some
increase.

The above are the direct deductions from the statistics. Be-
fore accepting these as final results, however, we must go back of
the statistics and inquire whether there are any circumstances
or considerations relating to the data on which these statistics
are based tending to modify the results. One such consideration
should be mentioned. With the development of medical science
there has been a gradual increase in correctness of diagnosis of
the cause of death. Professor Walter F. Willcox (2) showed in
1917 that a very considerable apparent increase in cancer mor-
tality would result simply from the continued improvement
in the correctness of diagnosis, which increase in correctness had
probably continued up to the present day. Undoubtedly in the
past many deaths attributed to old age should properly have
been set down as cancer, and many others attributed to other
causes would undoubtedly have been attributed to cancer if the
correct diagnosis had been made. The result of this would be

256

WENDELL M. STRONG

that if there were a really stationary cancer mortality it would,
nevertheless, appear to be increasing considerably because of the
increasing correctness of diagnosis.

How effective this could be to cause an apparent increase
where no real one exists may be seen in the statistics of appendi-
citis mortality which show an increase of 40 per cent from 1900 to
1915; yet undoubtedly the real rate was decreasing in this period
because of surgical advance and readiness to resort to surgery.

Moreover, increasing correctness in diagnosis works almost
altogether in one way; that is, it results in attributing to cancer
many deaths which would formerly have been incorrectly attrib-
uted to some other cause, and in changing but few the other
way. The effect of increased correctness in diagnosis, which is
undoubtedly more important at the advanced ages, would seem to
me enough to explain what apparent increase in cancer mortality
there is in ages above 65, and it would emphasize the fact that
for ages below 65 there has not been even an apparent increase.

Meanwhile the statistics of the United States Registration
Area show a continuous increase in cancer deaths. These are
population statistics, however, and are not analyzed as to ages;
hence, as compared to such statistics as we have considered, they
have very little meaning, since a change in the age proportion
of the population might be responsible for the entire apparent
cancer increase.

My conclusion would be, even taking into account the regis-
tration area's apparent increase, that we cannot now determine
whether the cancer mortality is slightly increasing, practically
stationary, or slightly decreasing, but that we can be sure it is
not greatly increasing. A more exact result is something for
future investigations when reliable statistics for a long period of
years are obtainable.

Lest what I have said be misinterpreted I would add that such
a conclusion does not lessen at all the seriousness of the cancer
problem. It merely holds out hope that the terrible scourge will
not increase without limit.

REFERENCES

(1) Dublin, L. I: Mortality Statistics of Cancer Among Wage Earners, etc., J.

Cancer Res., 1919, iv, 235.

(2) Willcox, W. P: On the Alleged Increase of Cancer, Trs. Am. Statistical

Assoc., 1917, xr, 701 and J. Cancer Res., 1917, ii, 267.

CANCER AND PARASITE

ISIDOR KROSS1

From Columbia University, Institute of Cancer Research, F. C. Wood, Director
Received for publication May 1, 1922

Ever since the first recognition of the bacteria as a cause of
disease, attempts have been made to demonstrate a bacterial
cause for malignant tumors. The importance of such a dis-
covery, with all its possibilities of prophylaxis or even cure, is
obvious, and accounts for the sustained interest in this question,
in the face of constantly repeated failure.

One of the most recent investigators is Nuzum (1) who em-
ployed a mouse carcinoma (no. 11) from the Crocker Institute.
From this growth he isolated a diplococcus and described a
malignant tumor having all the characteristics of carcinoma 11,
which followed inoculation of a culture of this organism into mice.

The importance of such a discovery would be, needless to say,
beyond computation, and it was therefore determined to repeat
the experiments with additional checks and controls. That
microorganisms are frequently found in malignant tumors, espe-
cially those growing in mice, is a well established fact, but, as
has been said, no one has ever succeeded in proving that they
play any role as causative agents.

To summarize Nuzum's findings, it will suffice to say that he
cultivated pieces of tumor in tissue ascitic fluid media covered by
paraffin, and thus partially anaerobic. From this culture he
isolated a Gram-positive diplococcus, which he considers the
specific bacterium in the production of carcinoma 11. A sub-
culture was made of this organism in the same manner, and
inoculated into mice. In three of them, the typical carcinoma 11

1 Adjunct Gynecologist, Mt. Sinai Hospital, New York.

257

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 4

258

ISIDOR KROSS

is said to have developed. However, careful perusal of the article
suggests that only two of these mice had a real tumor.

As can readily be seen, two factors must be considered in an
experiment of this nature: (1) the tumor cell proper; and (2) the
microorganism, if any. In order to test the ability of the latter
to produce the malignant growth in question, it is necessary to
prevent any participation of the tumor cells in the process. To
do this without affecting the microorganism, two methods were
employed in my own experiments, radiation and freezing.

In the first, several mice that had well developed carcinoma 11
tumors, were exposed to the Roentgen ray, being given a dose
known to destroy the tumor cells completely. The rayed tumor
was then excised aseptically and inoculated into 24 mice in the
usual fashion. Part of this same tumor was cultured, following
most minutely the technic laid down by Nuzum. As a control,
carcinoma of the same series, from another mouse, was inocu-
lated into 24 mice, and fragments were put at the same time
into twelve tubes of the ascitic tumor fluid media. The tumor
used in this control series was not radiated.

Transplantation of the unradiated control tumor resulted in
the usual 60 per cent of takes which this tumor has given during
the past seven or eight years at the Crocker Laboratory. On
the other hand, in not a single instance did a tumor develop
from the radiated cells. In other words, where the tumor-cell
factor was eliminated the microorganism factor, which was being
tested, was not able to reproduce the tumor.

Because of the extreme importance of the question, this experi-
ment was repeated in another series of mice, with the same nega-
tive results.

The cultures made from the radiated tumor were incubated
for five days and were then inoculated into 60 mice. In none
of the 60 did a tumor subsequently develop. At the same time,
smears were made from the bottom of the culture tubes and
stained with Gram's and with Giemsa's stain; these contained in
most instances many short and long bacilli, streptococci, staphylo-
cocci, and diplococci. In some smears all these varieties were
present and most of the smears, in fact, showed more than one

CANCER AND PARASITE

259

organism. In no instance was a pure culture of any one organ-
ism found.

The second method employed to eliminate the tumor cells
was alternate freezing and thawing of the tumor before inocu-
lation. The growth was removed aseptically and placed in a
sterile dish, where it was emulsified with scissors, and the dish
was then placed in an ice and salt mixture. After about fifteen
minutes, when the mass was frozen solid, it was ground with a
sterile pestle until it became soft and fluid. This freezing and
thawing was repeated twice, with the purpose of destroying the
tumor cells without harming any bacteria that might be present.
Part of this emulsion was inoculated into 36 mice, and another
part was cultured in the ascitic tissue fluid media.

In only two of the 36 mice was there any growth, and in both
instances the tumor was not discoverable until the twenty-
eighth day after inoculation, whereas in routine laboratory inocu-
lation, growth is generally noted in from seven to ten days.
Hence the freezing and thawing had killed all but a few of the
tumor cells. If carcinoma 11 were due to a microorganism, all,
or at least most of the mice, should have had tumors, both in
the frozen and thawed tumor series, in the rayed tumor series,
and in the culture inoculation series. The absence of a tumor
in all but two late instances, militates against the assumption
that carcinoma no. 11 can be ascribed to a microorganism, espe-
cially in view of the fact that no one type of organism was found
as a constant occurrence. The two growths that did develop
from frozen and ground tumor can be readily explained by the
assumption that the freezing did not destroy quite all the tumor
cells. Others of the staff at the Crocker Laboratory have had
the same experience.

In all the cultures the results, as concerns the specific organ-
ism described by Nuzum, were negative as has been already
stated and subcultures inoculated into 24 mice produced no
tumors.

A careful perusal of the experiments of Nuzum leads to the
conclusion that in only two instances out of eighty-nine did the
culture fluid inoculation cause tumor growth, i.e., animal no.

260

ISIDOR KROSS

250 in experiment 1, and a second animal in experiment 3.
These two instances and our own negative findings repeatedly
and carefully checked up suggest that in all probability the
tumors in Nuzum's experiments were spontaneous new growths.

It may be stated in conclusion that all the animals showing
negative results were kept under observation for a period of
three and a half months, so that ample time was given for a
tumor to arise.

CONCLUSIONS

1. While most of the ascitic tissue fluid cultures of Crocker
Institute carcinoma 11 contained microorganisms of various
kinds, in no instance was there found the characteristic micro-
organism described by Nuzum.

2. In not a single instance was the inoculation of mice with
these cultures followed by tumor growth.

3. It is probable that the two undoubted tumors in Nuzum's
series were a spontaneous new growth.

REFERENCE
(1) Nuzum: Surg. Gynec. andObst., 1921, xxxiii, 167.

TERATOMAS AND THEIR RELATION TO AGE

H. E. HIMWICH

From Department of Pathology, Cornell University Medical College, New York
Received for publication May 10, 1922

The great accumulation of reported cases of teratomas offers
an excellent opportunity to review the data, with the object of
investigating the relation of their frequency to the age of the
host. In the course of the study, it has become apparent that
this relationship is so definite as to assume the form of a gen-
eral law.

The largest collections of teratomas were gathered by Taruffi
and Ahlfeld. Both of these observers stressed the large con-
genital forms which are situated in the head, the thoracic, the
abdominal, or the sacral regions. Gonadal teratomas occur
most frequently in early adult life (Wilms). Among others,
Askanazy investigated the internal craniopagi, and Ekehorn,
the internal thoracopagi. Lexer and Nakayama studied the
abdominal inclusions and pygopagi.

Though many hypotheses have been advanced on the origin
of teratomas, they may be resolved into two view points: The
teratoma is either the offspring or the twin of its host. Stockard
has recently produced experimental evidence in favor of the
latter conception.

CRITERIA AND METHODS OF STUDY

Two precautions have been observed in compiling the present
statistics: (1) Only growths of tridermal or bidermal origin
have been considered. No such case has been omitted. (2)
Special effort has been exerted to determine the age at which
the teratoma began its growth. The first increase in size of
external growths may be accurately observed. For those

261

262

H. E. HIMWICH

situated internally, the initial symptom was used as an indicator.
Where the history was deficient, the age at which the operation
took place or at which death occurred was taken as the closest
approximation obtainable. The last criterium particularly
applies to the teratomas of the aged. In this manner 895 cases
have been studied. Sometimes a period of slow growth is
followed by one of heightened activity. Such is the case of
chorioma testis reported by Jackson, in which growth commenced
at the age of twenty and slowly continued to twenty-three,
after which the increase in size became extremely rapid. Since
the relationship between the growth of the host and that of the
teratoma is of interest, in such cases, the beginnings of both
periods have been noted. A similar effort was made for internal
teratomas, thus bringing the total number of growths tabulated
to 975.

In systematizing the results it was noted that though the
variation from year to year is considerable, there seems to be
an orderly waxing and waning of the number of cases to an
extent which justifies the drawing of a curve. An average has
been drawn in order to minimize accidental variation. Six
year periods have been chosen because they are the longest
which correspond to actual changes throughout the length of
the curve. The first period begins at fertilization and ends at
five. In the curves which are drawn to a scale of one-half,
the abscissae represent the age of the host when the tumor began
its growth, the ordinates the number of cases in each year.

THE CURVE FOR ALL TERATOMAS

After an initial maximum rise, the curve falls and remains
low from five to eleven years (fig. 1). At eleven it achieves a
higher level, which is increased at seventeen and twenty-three
years. The second maximum is found between twenty-three
and twenty-nine years. The curve falls gradually at twenty-nine
and thirty-five years, and then more rapidly at forty-one, after
which it becomes progressively lower towards its end at seventy-
six years. Eighty-seven per cent of the teratomas occur before
forty-one years and 95 per cent before fifty- three.

TERATOMAS AND THEIR RELATION TO AGE 263

Most congenital teratomas do not evince postnatal growth and
since this study concerns only those which do grow, for the early
maximal total may be substituted the smaller number of tumors,
showing power for growth, represented by the lower broken
line in the graph. With this correction the highest point in the
curve is found between twenty-three and twenty-nine years.
This is borne out by Wilms who finds the period of greatest
frequency for sex-gland teratomas to be between the ages of
twenty and thirty years. His conclusion is to be expected
because teratomas occur most frequently in sex glands.

Fig. 1. Curve for all Teratomas

The abscissae stand for the years of appearance of growths, the ordinates for
the number in each year. Both are drawn to a scale of one-half. The total
number of congenital tumors is not shown. The dash line indicates the average
for each six-year period. In the first six years there are two; the upper one
stands for all cases, the lower for those which had power of postnatal growth.
The mode is from twenty-three to twenty-nine, the time when growth stops.

STUDY OF THE CORRECTED CURVE

There are three aspects of the described phenomenon: First,
it is evident that the total number of teratomas at any age
increases with the actual growth of the individual. As size

264

H. E. HIMWICH

increases, the total number of teratomas increases. But it
must be remembered that while size is increasing, growth rate
is falling. Therefore, second, teratomas become more frequent
as growth slows down, at the time when growth potential be-
comes smaller. Growth must be recognized as involving two
elements: increasing actual proportions, and decreasing growth
potential. Hence, third, the total number of teratomas at any
age increases as growth potential diminishes.

When one recalls the fact that the changes in growth rate
are not constant, the number of teratomas is seen to bear even a
closer relationship to the growth of the hosts than has been
indicated. Not only do these tumors appear as growth of the
host slows, but during their time of appearance the teratomas
are more frequent in the periods of slower growth of the host.
Teratomas are common in early infancy following a space of the
most rapid proliferation of all — fetal growth. They increase
again when the comparatively rapid growth rate of early child-
hood gives way to the slower one of pubescence and lastly they
are found in greater numbers as active growth gradually ceases.

To explain the relative number of teratomas appearing in the
several periods we must take into consideration an additional
factor, the growth potential of the embryonal rest, for a teratoma
by most theories arises in an embryonal rest of some kind. It
is known that the great majority of embryonal rests do not grow;
they either degenerate or remain dormant. Others achieve a
more or less perfect adult growth, while a few develop into tumors.
Tridermal rests act in a similar manner. Thus we have a few
with high growth potential, many with less capacity for growth,
and finally others which remain latent unless they are stirred
to development by an external stimulus.

During the first six years of life the growth rate of the host
declines rapidly, his growth potential is greatly reduced, and
the number of teratomas of relatively higher potential, capable
of proliferation, is comparatively large. From five to eleven
years the growth rate is fairly constant, the loss of potential is
small, and the additional number of teratomas released is few.
The next appreciable change takes place during the differentia-

TERATOMAS AND THEIR RELATION TO AGE 265

tion occurring at puberty and accompanying the maturation
of the sex organs. Here the number of teratomas begins to
increase and continues to do so until growth finally stops, at
which time the greatest number of teratomas make their pres-
ence known. Although the loss of growth potential in the host
becomes smaller, being least in the final period from twenty-
three to twenty-nine years, yet just because growth itself is
slow, there is an ever increasing number of embryonal rests of
low potential, capable of expressing their latent growth energy.
The very inactive rests even at this period do not have sufficient
energy to start growth spontaneously and are therefore consid-
ered in another group at a later time. However, they may serve
as a nidus for neoplasms since many do commence development
after the growth of the host has stopped.

Thus the number of teratomas appearing at any given time
depends upon the amount of loss in growth potential of the
hosts during that time and the number of rests whose potential
is large enough to proliferate under these conditions. Evidently
the number of teratomas in any period varies inversely to the growth
potential of the hosts and directly as that of the embryonal rests.

ANALYSIS OF CURVES OF AGE INCIDENCE OF TERATOMAS IN THE
DIFFERENT LOCATIONS

The object of this study is to show the relation to the general
law of the occurrence of teratomas in the various situations.

Thoracopagi. Ekehorn in his collection of teratomas of the
anterior mediastinum finds fifteen cases occurring between the
ages of twenty and thirty, four between thirty and forty years,
and four more to sixty, thus agreeing with the present curve in
showing the maximum occurrence from twenty-three years to
twenty-nine.

Abdominal inclusions. The abdominal tridermal growths
are frequently discovered at birth but continue to be found
throughout life, with a second rise from twenty-three to twenty-
nine years.

Craniopagi and pygopagi. There is not a sufficiently large
number of head and sacral teratomas to yield reliable statistics.

266

H. E. HIMWICH

Statistics of various authors. They show a close agreement with original statistics
here presented. Note increased number of dermoids during pubescence

thoracopagi
(christian)

DERMOIDS
(PAULl)

EMBRTOMES
(CHEVAS8U)

8EMINOMES
(CHEVASSU)

RCINOMA

Testis

Ovary

10-00

X o

0-^

0-19

0-5

90-30

Q
o

20-25

5-11

30-40

Ivy

18-94

J.U

25-30

11-17

40-50

15-20

25-29

30-35

17-23

50 +

20-25

30-34

35-40

23-29

25-30

35-39

40-45

29-35

30-35

40-44

45-50

35-41

35-40

45-49

50-60

41-47

40-45

50-59

60-63

47-53

45-50

60-75

53-59

50-55

59-65

55-60

65-71

60-65

65-70

70 +

103

157

207

Inclusions according to age and situation. Those which are capable of growth appear
in greater number from twenty-three years to twenty-nine years

AGE

CRANIOPAGI

THORAC-
OPAGI

ABDOMINAL
PARASITES

OVARIAN
TERATOMAS

TESTICULAR
TERATOMAS

PTGOPAGI

TOTAL

-5

279

5-11

11-17

17-23

114

23-29

144

29-35

116

35-41

41-47

47-53

53-59

59-65

65-71

71-77

Total. . . .

123

132

313

264

975

TEE ATOM AS AND THEIR RELATION TO AGE

267

Most are congenital. The internal craniopagi occur most
frequently from eleven to twenty-seven years. The only year
where there is more than one case is the nineteenth, where
there are two. The last teratoma capable of spontaneous growth
appeared at twenty-seven. The later examples of sacral tera-
tomas are even rarer than those occurring in the head. There
is one each at ten and thirteen years and four from nineteen to
twenty-three.

Ovarian teratomas. Turning next to the gonadal teratomas
and comparing the testicular and ovarian curves, we see that
the latter is less variable (fig. 3). This is due to the delayed
diagnosis of so many of the ovarian teratomas, probably because
of their slow growth, as their structure is often of the adult type.
Their internal position further postpones their discovery, yet
dermoids are sometimes found by accident. Nevertheless,
the largest number of tumors occurs between the ages of twenty-
three and twenty-nine. From a review of one hundred and
three cases Pauli finds dermoids appearing most often from
twenty to thirty years.

Another difference between the ovarian and testicular curves
is the greater rise in the former in the two periods between
eleven and twenty-three years. This phenomenon might be
expected as a result of the growth differences since, in the male,
postpubescent growth is more rapid than in the female.

Teratoma testis. The external situation of the male sex gland
allows prompt discovery of its tumors. Although the greatest
decline in rate of growth occurs early in life from birth to four
years and the number of testicular teratomas at this time is
large, the maximum number nevertheless occurs at a later period.
This is probably due to the fact that growth under four years is
comparatively very rapid in spite of its fast declining rate. The
modal year of the curve is twenty-six. It is interesting to note that
the last growth cartilage of the long bones ossifies at twenty-five.

Teratomas in the testes commence growth most frequently
between the ages of twenty- three and twenty-nine years. Che-
vassu finds the maximum between twenty and thirty years
(%. 2).

268

H. E. HIMWICH

TERATOMAS OF LATER LIFE

After establishing the time of greatest frequency of teratomas,
it still remains to account for those of old age. Those diagnosed
after the age of fifty-one may be divided into two groups. The
first consists of neoplasms of adult structure which had reached
the limits of their capacity for post-natal growth, while those of
the second are more malignant. To the first group belong such
tumors as the following:

Craniopagi. Beck reports a case in which a dermoid was
found at autopsy in place of the hypophysis in a woman seventy-
four years of age. Eberth reports a similar accidental finding
beneath the dura in a woman of seventy-five.

Thoracopagi. There are two examples in Ekehorn's collec-
tion— Pinder's case of a patient with bulbar paralysis, aged fifty-
three, in whom the dermoid was discovered at autopsy; and
Lebert's of a man of sixty who had been dyspneic since his six-
teenth year.

Abdominal inclusions. Rizzoli (Taruffi) reports two cases
of late abdominal inclusions, one at sixty, the other at sixty-two.
Symptoms had been present for a long time in both. In one
of the cases they appeared first at the age of twenty.

The tumors mentioned thus far were benign, though some
produced symptoms because of their size and position.

In the seeond group of neoplasms the element of trauma be-
comes important in the etiology. There is Bonney's report of a
retroperitoneal chorioma of a man of sixty-seven, and GoebelPs
of an abdominal teratoma that became malignant at fifty-four,
twenty-seven years after a mass had been diagnosed. Djewitski
reports a chorioma of the bladder which first gave symptoms
at the age of seventy-three. The same irritation which produces
a papilloma of the bladder may transform an otherwise benign
embryonal rest.

Pygopagi. Hudson describes a sacral teratoma which began
growth at the age of fifty-two; the history shows that a nodule
had existed in that region since the birth of the patient. The
histological picture is one of a tridermal rest with cancerous
degeneration of mucous glands. It is similar to an old age

TEKATOMAS AND THEIR RELATION TO AGE 269

cancer arising in previously normal tissue. Evidently in the
last four cases it is not growth potential of the embryonal rest
but an extrinsic traumatic influence, to which every part of the
body is subject, that caused the proliferation of cells. A case
of Briddon's beautifully illustrates both these factors occurring
in the same growth but independently and at different times. It
concerns a sacral dermoid which appeared externally at the age
of twenty-two and then ceased growth till the fifty-second year,
when it underwent epitheliomatous change.

Teratoma testis. There are three examples of late teratoma
testis. Lexer quotes one from v. Bergmann's clinic in a man of
sixty. On section the growth was of adult structure.

Ewing and Pepere report cases which first showed growth at
the ages of sixty-one and sixty-three, respectively. The micro-
scopic examination in both instances showed carcinomatous
change of one element in a totipotent rest.

Ovarian teratomas. Of thirty-two dermoids, thirteen exhibited
malignant transformation of a carcinomatous, sarcomatous, or
endotheliomatous type. Four showed thyroid structure, of
which three were rapidly growing tumors. In six reports de-
tails were lacking. However, since the growths were called
dermoids, their structure must have been of the adult type,
like that of the remaining nine inclusions.

To summarize, the late appearance of stationary teratomas is
due to their delayed discovery, while that of growing teratomas
is caused by their injury.

Carcinoma testis. Carcinoma testis is discussed in this place
not only for its possible teratomatous origin, but because trau-
matic etiology links it with the tumors of later life. In many cases
of teratoma testis in young people the transformation of a slowly
into a more rapidly growing tumor is caused by trauma. In
older people the growth is rapid from the start. The same se-
quence of events obtains for carcinoma testis. The cell of
many cases of carcinoma testis is characteristic, with a large
nucleus and clear cytoplasm. Sometimes the growth is called
a sarcoma; the difference in opinion is due to the fact that no
analogous cell is found in the human body. This same cell is

270

H. E. HIMWICH

often found with teratoma testis. There are only two probable
interpretations: (1) The irritation caused by some extrinsic
factor, in this special case, by the teratoma on the tubule cells,
is the cause of carcinoma. (2) The unique type of cell is of
teratomatous origin. Chevassu takes the position that it
develops from the adult spermatogonia, putting the tumor in
the class of acquired carcinomas. The final convincing link
in the chain of evidence has not been produced, for he has not
been able to trace the steps of anaplastic change from the sperma-
togonia to the carcinoma cell.

It has been definitely established that the cell which is of more
rapid growth will often overrun and may finally crowd out alto-
gether the other constituents of the tumor. Thus arise the
rhabdomyomas, the chondromas — the simple tumors of the
sex glands. In this uncontrolled competition the most embryonal
type of cell would have a decided advantage. Therefore Ewing
concludes that carcinoma testis is a one-sided teratoma. In
the light of the foregoing it is interesting to see to which of these
two theories the carcinoma curve lends itself.

According to Chevassu's statistics embryomas occur with
greatest frequency from twenty-five to thirty years and semino-
mas from thirty-five to forty (fig. 2). The writer's review of a
larger number of cases coincides with the data of Chevassu, the
modes occuring from twenty-three to twenty-nine years in the
teratomas and thirty-five to forty-one years in the carcinomas.

Comparing the carcinoma testis curve with that of all cancer
(Hoffman1) , of which the congenital cases are too small a propor-

1 Mortality from cancer throughout the United States Registration Area. All
organs and all parts. 1903-1912. (Hoffman, The Mortality from Cancer.)

MALE

FEMALE

Until 10

1,170
2, 028
3,757
10, 750
24, 431
35,327
33, 745
18,3S1

984
1,844
7,891
26, 779
46, 669
52,393
43,010
24, 601

10-24
25-34
35-44
45-54
55-64
65^74

75 and over

TERATOMAS AND THEIR RELATION TO AGE

271

tion materially to alter the general outline, we see that the former
has no resemblance to the latter, for in that case it would have
a continuous rise to some time after sixty. In brief, the carcinoma
testis curve is the teratoma testis curve with the mode slightly
shifted.

Since the histogenesis of carcinoma testis has not been traced
from either embryonal or adult tubule cells, it is probable that
carcinoma testis is of nontesticular origin, and since there is no
reason why misplaced cells should so often be of the same type
or occur so frequently with teratomas, unless they are of tera-

e-t J'-'t. 34-39 USf iO-if ~"?e^7 >

Fig. 2. Testicular Teratomas

The lower two curves are reproduced from Chevassu's paper. The con-
tinuous line represents the teratomas; the dash line, the carcinomas. In order
better to compare the new curves with those of Chevassu, the number of tera-
tomas in each year was divided by four, and that of carcinomas by two. Both
curves are plotted on a basis of ten-year periods.

tomatous origin, we are forced to this conclusion towards which
the study of the curve gives additional evidence. Similarly to
teratoma testis, carcinoma may show a congenital increase in
the size of the organ. Morestin (Chevassu) reports such a tumor,
which assumed malignancy at the age of thirty-seven. Like-

272

H. E. HIMWICH

wise carcinoma occurs more often in undescended testicles. It
begins its growth in the rete and, like teratoma testis, is occa-
sionally observed in pseudohermaphrodites.

Ovarian tumors. The origin of primary tumors of the ovary
is so undecided that any data in reference to them is of particular
interest. Here we shall mention a few facts in regard to one of
these tumors which may be of teratomatous origin, i.e., the
sarcoma.

Ewing divides these sarcomas into three main types: (1)
spindle cell; (2) round cell; (3) myxoma cell. This classification
is of special significance since just such types of sarcomatous de-
generation of dermoids have been observed (Debucy).

e S " '» « at 35 v/. * > S3 sf 1/ 77

Fig. 3. Ovarian Teratomas

These curves are averages for six year periods drawn to a scale of one-half.
The continuous line represents the teratomas with the mode from twenty-three
to twenty-nine years; the dash line the carcinomas with the mode from eleven to
seventeen years, the period of pubescence.

Desurmont finds that the different primary tumors are bilateral
to varying degrees. However, sarcomas, 25 per cent, and der-
moids, 20 per cent (Pauli), approximate each other quite closely.

Finally, the most common tumor of infancy is the sarcoma,
which is most frequent at fifteen years (Donhauser). Cordier
and Zangemeister give fifteen and twenty years, respectively,
as the age of most common occurrence of sarcomas. They are
found from fifteen years to twenty-five and from forty years to
fifty (Desurmont), both periods of physiological stimulation.
The writer finds the mode of the combined carcinoma and sar-
coma curve at fifteen years (fig. 3). Comparing the mode of
this curve with that of ovarian teratomas, we find that it has
been shifted forward to the time of pubescence. Hence there is
a group of embryonal cell tumors having an age incidence simi-
lar to teratomas, and becoming malignant under the stimulation
of puberty.

TERATOMAS AND THEIR RELATION TO AGE 273
FACTORS IN THE ETIOLOGY OF TERATOMAS

The growth of the earlier teratomas may be adequately ex-
plained on the basis of a growth competition between the host
and the embryonal rest. But even in the teratomas of infancy
another factor, trauma, may be present. It becomes increasingly
important later on.

Growth potential. The growth of the host inhibits that of the
teratoma.

The growth of the embryonal rest may be divided into
two parts : its prenatal development, or growth which continues
until stopped by the inhibition produced by the excessive
growth of the host; and the growth of which it is still
capable (growth potential) after that of the host slows down
or ceases. These two parts are in reciprocal relation to each
other. The earlier the prenatal inhibition, the smaller and less
differentiated will be the inhibited rest but the greater will be
the remaining growth potential. Small embryonal rests may
develop proliferative powers while large ones, which achieve a
certain intrauterine development, seldom if ever show further
capacity for spontaneous growth.

Trauma. The shift in the mode of the curve of carcinoma
testis from twenty-three to twenty-nine years to thirty-five to
forty-one years is due to an external stimulus.

In spite of the larger growth potential of the smaller testicular
rests many do not achieve malignancy until their immediate
region is traumatized. This is illustrated on comparing these
tumors in the testicle and the ovary. A larger proportion of
the latter are benign, forming adult structures. The chief
difference in their histories is due to their locations; the testicle
is exposed to injury, the ovary is not. It is generally admitted
that teratomas do not become malignant much oftener than do
normal tissues.

In the case of carcinoma testis we should expect with a history
of injury to a small undifferentiated rest and the resulting pro-
liferation of an embryonal cell, a shift in the mode of the curve
towards that of old age cancer.

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 4

274

H. E. HIMWICH

SUMMAEY AND CONCLUSIONS

1. Growth 'potential and teratomas

1. Parasite. Teratomas are tridermal embryonal rests en-
dowed with a certain amount of possible growth, i.e., growth
potential. When the rest is comparatively large it has necessarily
consumed considerable growth energy before birth, while in
the small teratoma, the growth period may be divided into two
parts, a slight early growth soon followed by an inhibition, and a
later, or post-natal growth, should conditions permit.

2. Host. The total number of teratomas in a population,
up to any given age, increases while growth potential of the
hosts decreases. As the larger increases in the number of tera-
tomas occur in periods when growth of the host is slowed most,
the growth of the host must inhibit that of the embryonal rest.

3. The number of teratomas appearing in any given time
varies inversely with the growth potential of the host and directly
as that of the embryonal rest. The tumors which begin their
postnatal growth before that of the host stops are of highest
potential, but are not necessarily more malignant, for they must
overcome a still present inhibition. Since most teratomas have
a low growth potential, they appear most commonly at the
time the growth of the host stops — from twenty-three to twenty-
nine years.

77. Trauma and autonomous growths

1. Teratomas which start growth as a result of injury are
malignant more frequently than those which proliferate solely
under the influence of growth potential.

2. ^Tien trauma precipitates growth, the teratoma is fre-
quently monodermal. If the inclusion is still in an undifferen-
tiated condition the cell is often of an embryonal type. If a
developed inclusion is traumatized the cell in many instances
is like that of acquired cancer.

3. The curve of carcinoma testis rises and falls in a manner
similar to that of teratoma testis and not like that of old age

TERATOMAS AND THEIR RELATION TO AGE 275

cancer. This is another fact which may be adduced in support
of the theory that carcinoma testis is a one-sided teratoma. In
the female, a similar neoplasm might be expected to arise as the
result of the physiologic stimulations of puberty. This is what
actually takes place, hence the growth, in all probability, is of
teratomatous origin.

4. Trauma is followed by proliferation of cells, and any pre-
cipitant of regeneration may be important in the etiology of
acquired cancer. In the old the inhibition of the organism is
almost negligible. Hence trauma at that time may readily be
followed by an uncontrolled and therefore excessive growth.
Thus, loss of growth restraint may be almost as important a
factor in the etiology of acquired cancer as in that of congenital
inclusions.

REFERENCES2

Barron: Teratomata of the brain, Jour. Cancer Res., 1916, i, 311.
Christian: Dermoid cysts and teratomata of the anterior mediastinum, Jour.

Med. Res., 1902, vii, 54.
Desurmont: Etude anatomo-clinique des tumeurs solides bilaterales des ovares,

These de Paris, 1911-1912.
Ewinq: Neoplastic Diseases, Philadelphia, 1919.

Ewing: Teratoma testis and its derivatives, Surg., Gynec. & Obst., 1911, xii, 230.
Lexer: Ueber teratoide Geschwulste in der Bauchhohle und deren Operation,

Arch. f. klin. Chir., 1900, lxi, 648.
Nakayama: Ueber Congenitale Sacraltumoren, Arch. f. Entwcklngsmechn. d.

Organ., 1905, xix, 475.
Stockard: Developmental rate and structural expression, Am. Jour. Anat.,

1920-1921, xxviii, 115.
Taruffi: Storia della Teratologia, 1881-1894.

2 These references were chosen out of over one hundred and sixty; they contain
extensive reviews of the literature consulted.

AN ATYPICAL ADENOMA OF THE PANCREAS ORIGI-
NATING IN ISLET TISSUE

HARRY GOLDBLATT

From the Department of Pathology, School of Medicine, Western Reserve University,

Cleveland, Ohio

Received for publication May 13, 1922

True simple adenomata of the pancreas arising from the
islands of Langerhans have been described by several authors.
The condition is very rare and the diagnosis is somewhat difficult
to establish. The tumors hitherto described have been very
small, the largest being 11 mm. in diameter. Most of them were
accidental autopsy findings, and some were discovered only in
the sections prepared for routine microscopical examination.
Nichols (10), in 1902, was one of the first to report such a tumor.
He stated clearly and convincingly upon what grounds he based
the diagnosis, and subsequent authors seem to have guided
themselves according to those facts in deciding the islet origin
of an adenoma of the pancreas as opposed to the other two pos-
sible sources, viz., duct and adult acini. Helmholz (3) in 1907,
Morse (8) in 1908, and Koch (6) in 1914, described cases which
fulfilled the requirements stated by Nichols and concluded that
they were dealing with a simple adenoma originating in the
islands of Langerhans. LeComte (7) in 1913, described a case
of his own and also reviewed the literature to that date. He
determined that adenomata of the pancreas have no great clinical
significance and " constitute no very serious hindrance to long
life.'^ He classified as true adenomata of the islands those
reported by Nichols (10), Ssobolew (13), Reitmann (11), Herx-
heimer (4), Helmholz (3), Morse (8), Cecil (1), Weichsel-
baum (14), Rollet (12), and Heiberg (2). Although he favors a
diagnosis of adenoma in the above cases, some of the authors
themselves, notably Ssobolew, Herxheimer, Weichselbaum, and

277

278

HAKRY GOLDBLATT

Cecil, either consider the nodules to be hypertrophic islands, or
do not decide definitely between a diagnosis of simple hyper-
trophy and neoplasm. Whenever there is doubt between the
diagnosis of simple hypertrophy and new growth he recommends
the latter if the nodule exceeds the arbitrary diameter of 1 mm.,
provided, of course, that it shows all the characteristics of island
tissue. Others have reported the same lesion and several have
gone so far as to trace the origin of some carcinomata of the
pancreas to the islands. The conclusions of some of these
authors, especially of Fabozzi, who reported five cases, have
been disputed vigorously. With the exception of Horgan (5)
most of these investigators studied advanced carcinomata so
that a doubt of validity of their conclusions seems justifiable.

The tumor of the pancreas to be described here was discovered
at autopsy by Dr. H. T. Karsner, performed through the courtesy
of Dr. John Maclachlan.

SUMMARY OF CLINICAL HISTORY AND PHYSICAL EXAMINATION

The patient is an unmarried white female fifty-five years of
age, in whose past history nothing of importance is noted save
that eight years ago a myomectomy was performed with un-
complicated recovery. Two years ago she noticed looseness of
the bowels, which over several months progressively became
worse and ultimately became a profound diarrhea. This was
accompanied by gradual loss of flesh and strength but she was
not confined to bed until about two months before death. In
the last weeks of her life bowel movements were extremely
frequent but well controlled, sometimes numbering 16 to 20 in
the course of twenty-four hours. No microscopic or chemical
examination of the stools was made but the gross examination
showed thin watery stools with very little mucus and no blood.
At various times indigestion of the different food elements was
noted. On a milk diet the stools, upon standing, showed a
supernatant layer of undigested fat. When meat was introduced
into the diet, undigested fragments were seen in the stools.
Fruits such as banana were easily recognizable. Fermention was
never marked nor was putrefaction severe. The clinical exami-

ATYPICAL ADENOMA OF PANCTIEAS

279

nation showed no abnormality of lungs, heart, abdomen, or
nervous system. The patient finally died apparently as the
result of profound asthenia.

SUMMARY OF AUTOPSY FINDINGS

The body is that of a white female apparently fifty-five years
of age, showing moderate emaciation. Upon examination of the
internal organs there were found, in addition to the tumor of the
pancreas described in detail below, slight chronic interstitial
nephritis, fatty metamorphosis of the liver, subacute entero-
colitis, and slight passive congestion of lungs, liver, kidneys,
and intestines.

SPECIAL EXAMINATION

Gross description

When the abdomen is opened the pancreas is found in the
normal location, but a tumor mass partly embedded in the head
presents above the lesser curvature of the stomach. This tumor
is not adherent to any of the structures in the neighborhood.
The pancreas measures 15 cm. in length. Partly embedded in
the head of it, 10 cm. from the tip of the tail, is a mass, roughly
spherical in shape and measuring 4.5 by 3.5 by 2.5 cm. This
tumor is definitely and completely encapsulated so that it can
be enucleated easily from the head of the pancreas. The capsule
is thin, pearly white in color, and in it, coursing over the surface
of the growth, are several dilated vessels, apparently veins. The
tumor is firm but elastic and cuts with moderate resistance.
The cut surface bulges moderately, is light yellow in color, and
bleeds slightly. There is moderate vascularization of the tissue
but no areas of hemorrhage are seen. Near the center of the
mass there is one small area of necrosis about 3 mm. in diameter.
The central portion of the growth is of a slightly deeper yellow
color, is somewhat firmer, and bulges to a greater extent than
the tissue at the periphery. The tissue is only moderately
friable. The head of the pancreas is normal save for moderate
congestion. The entire portion of the pancreas from the tumor
to the tip of the tail is the seat of marked atrophy. In this

280

HARRY GOLDBLATT

portion the cross section measures 12 by 5 mm. The lobules
are obviously reduced in size, there is slight increase of the
interlobular connective tissue, and a moderate amount of fat
infiltration.

HISTOLOGICAL DESCRIPTION

Sections were stained by hematoxylin and eosin, eosin meth-
ylene blue, thionin, and Mallory's connective tissue stain.

Pancreas

The duodenal portion shows very advanced autolytic changes,
so advanced, in fact that little description is justified beyond
noting a slight increase in connective tissue both between and
within the lobules. Several definite islets are found which are
of normal size and structure. The tail shows great reduction
in the size of the lobules, acini, and individual cells comprising
them. There is a great increase of the interlobular as well as
interacinar connective tissue, and in it are a few foci of lymphoid
cells. The smaller ducts are well preserved and normal, but
the larger ducts show desquamation of the lining epithelium in
some instances and complete digestion in others. The islets of
Langerhans are moderately abundant, of normal size and struc-
ture, and fairly well preserved. There is no sign of invasion of
the pancreatic tissue by new growth or of neoplastic change in
the cells of the pancreas.

Tumor

There is a definite but thin capsule which consists of dense,
moderately nucleated fibrous connective tissue. This shows
no sign of invasion by the parenchyma of the tumor. Through-
out the tumor there are present anastomosing bands of con-
nective tissue, some of which are densely fibrous and some rather
loosely arranged. Many of these trabeculae do not join at all
and are seen as isolated masses of connective tissue which vary
considerably in size and in shape, some being elliptical or rounded,
and others stellate or irregularly branching. Nearly every one
of them contains one or more blood-vessels. Some of these

ATYPICAL

ADENOMA OF PANCREAS

281

vessels are very thin-walled, consisting apparently of only a
single layer of endothelial cells, while others are large and well
formed with a thick wall consisting of intima, media, and adven-
titia. Most of the vessels are filled with blood. In some of the
trabeculae there are seen merely blood spaces showing no definite
endothelial lining. The trabeculae give to the parenchyma of
the tumor an alveolated appearance. The parenchyma con-
sists of numerous masses and anastomosing bands and strands
of epithelial cells. These masses vary greatly in size and shape.

t*M i *** - w > * r'L -J-** J ft

ibT^k .Jv^j

Fig. 1. Photomicrograph Showing Type of Cell, Arrangement of Cells
and the Supporting Connective Tissue Containing Capillaries

Some are in intimate contact with the bands of coi nective tissue
described above, but in most instances they are retracted from
them. This is very likely due to fixation. Between the indi-
vidual epithelial cells there is no reticulum. Cell outline in
many areas is rather indistinct, but for the most part the cells
are polyhedral or rounded in shape. Most of the cells are ar-
ranged in the form of solid masses, but some of the latter are
hollowed out and in a few sections apparent attempts at acinar
arrangement are seen, the cells surrounding these imperfectly

282

HARRY GOLDBLATT

formed acini being of the cylindrical type. In many sections,
what at first appear to be fairly well-formed acini are seen to
contain in their lumen small masses of loosely arranged moder-
ately nucleated connective tissue. These are formed apparently
by the ensnaring of some of the connective tissue stroma by
strands of epithelial cells which happen to be arranged in a
circular manner. Many of these small areas of connective
tissue within the pseudo-acini contain one or more small capil-
laries. Wherever the connective tissue is small in amount the
epithelial cells come into such intimate contact with the capil-
laries that a spinous peritheliomatous appearance is the result.
It is impossible to identify a well formed typical acinus, but in
a few cases, owing to the cylindrical appearance of the cells and
the pseudo-acinar arrangement one cannot rule out the possibility
that this represents a feeble attempt at the formation of a tubule
or gland follicle.

Most of the tumor cells are polyhedral or rounded in shape,
are approximately of the size of islet cells, in most instances
definitely smaller than those of pancreatic acini, and show the
same variation in size that the islet cells do. An occasional
very large cell is seen but this is not uncommon in the normal
islet. The cytoplasm of the cells is relatively moderate in
amount, is definitely acidophilic, stains very lightly, and is
granular. As in the nuclei of the islet cells, the chromatin is
aggregated into a single large mass or several, usually three or
four, small clumps. The examination of very many sections
failed to reveal a single mitotic figure.

The general architecture of the tumor as described above
bears a very striking resemblance to that of the islets of Langer-
hans, so much so that if certain portions in which the masses
and cords of epithelial cells are not large were seen in a normal
pancreas they would fulfill all the morphological requirements of
an island. A careful study of the cells themselves confirms this
very st l iking resemblance.

SUMMARY

That this tumor is of benign nature is deduced from the fol-
lowing facts: It is definitely and completely encapsulated.

ATYPICAL ADENOMA OF PANCREAS

283

Neither the capsule nor the neighboring pancreatic tissue shows
any signs of direct invasion by the new growth. There is no
metastasis. There are no mitotic figures. The general arrange-
ment of the tumor is not unlike that of some cellular adenomata
found in other organs of the body. The epithelial cells of the
tumor and their arrangement with relation to the vascularized
trabeculae show unmistakable signs of differentiation, since there
is such a striking resemblance to the islets of Langerhans. In
size, shape, staining characteristics, and the minute structure of
the cytoplasm and nucleus the cells resemble very closely those
of the islands.

The tumor here described is in all respects like those reported
by Nichols, Helmholz, Morse, and others. The only differences
are that this growth is much larger than any hitherto reported,
and that by reason of its size and location it had exerted pressure
upon the pancreatic ducts and had caused clinical signs and
symptoms suggestive of pancreatic insufficiency. In this respect
it differed from most adenomata hitherto described. Glycosuria
was never found.

CONCLUSION

An adenoma of the head of the pancreas has been described
which originated in an islet of Langerhans.

REFERENCES

(1) Cecil, R. L. : Jour. Exper. M., 1911. xiii, 595.

(2) Heiberg, K. A.: Centralbl. f. allg. Path. u. path. Anat., 1911, xxii, 532.

(3) Helmholz, H. F. : Bull. Johns Hopkins Hosp., 1907, xviii, 185.

(4) Herxheimer, G. : Virchows Arch. f. path. Anat,, 1906, clxxxiii, 228.

(5) Horgax, E. J. : Jour. Lab. and Clin. M., 1920, v, 429.

(6) Koch, K. : Virchows Arch. f. path. Anat.. 1914, ccxvi, 25.

(7) LeComte, R. M.: Jour. Med. Res., 1913, xxix, 251.

(8) Morse, M. E.: Jour. Amer. Med. Assoc., 1908, li, 1075.

(9) Mussey, R. D. : Med. Clin. North America, 1919, iii, 681.

(10) Nichols, H. G. : Jour. Med. Res., 1902, viii, 385.

(11) Reitmann, K. : Ztschr. f. Heilk,. 1905. xxvi, Abt. f. path. Anat., 1.

(12) Rollet, H.: Frankfurter Ztschr. f. Path., 1912, x, 268.

(13) Ssobolew, L. W. : Virchows Arch. f. path. Anat., 1904, clxxvii, Suppl. 123.

(14) Weichselbaum, A. : Sitzungsbericht d. kaiserliche Akademie d. Wissen-

schaft. in Wien, Mathem. natur. Klasse, 1910, cxix, 73.

THE INFLUENCE UPON THE GROWTH OF TRANS-
PLANTED FLEXNER-JOBLING RAT CARCINOMA OF
HYDROGEN IONS AND OF VARIOUS SALTS IN DIF-
FERENT CONCENTRATIONS

KANEMATSU SUGIURA, HELEN MILLER NOYES, and K. GEORGE FALK

From the Huntington Fund for Cancer Research, Memorial Hospital, and the
Harriman Research Laboratory, The Roosevelt Hospital, New York

Received for publication May 16, 1922
1. INTRODUCTION

In connection with the investigation of certain enzyme actions
of extracts of malignant human and rat tumors, the results of
which are published elsewhere (1), a study was made of the effects
upon their subsequent growth of various treatments of the Flex-
ner-Jobling rat carcinoma transplants before inoculation. The
results presented here include the effects of solutions of different
hydrogen ion concentrations and of some simple salts and their
mixtures.

2. EXPERIMENTAL METHODS AND RESULTS

a. Tumor transplantation method

Albino rats, both male and female, were used in this work.
They were inoculated with the Flexner-Jobling rat carcinoma in
the usual way, and the tumors were allowed to grow for a period

FRC

of from six to seven weeks (2). The term for example,

indicates that the tumor was derived from the 86th generation,
series A. The rats were fed on wheat bread soaked in whole
milk, fresh cabbage or carrots, and fresh tap water ad libitum.

In carrying out the transplantations, six (or more) small
particles of tumor tissue each weighing about 5 mgm., and one
piece weighing about 50 mgm., were selected from the non-
285

286 K. SUGIURA, H. M. NOYES AND K. G. FALK

necrotic area of the rat tumor and placed in 25-cc. portions of
solutions of definite hydrogen ion concentration and salt con-
tent previously sterilized in a steam autoclave at fifteen pounds
pressure for fifteen minutes. The flasks containing the solutions
and tumor fragments were allowed to remain in the ice box for
definite intervals of time of from one-half to seventy-two hours.
At the end of this period of time, the smaller pieces of tumor
tissue were inoculated with a trochar into rats subcutaneously
in the region of the right axilla. At the same time, the larger
as well as some of the smaller pieces of tumor tissue were fixed
in formaldehyde solution for histological examination. As a
control, the same number of animals were inoculated with un-
treated tumor tissue immediately after its removal from the
tumor-bearing animal. The progress of the transplanted tumors
in the rats was recorded graphically by measuring them every
week by means of calipers.

b. Preparation of solutions

The pH values of the solutions were determined in most cases
by means of indicators and the standard solutions recommended
by Clark (3), and in some potentiometrically. The salts used
were purified by crystallization in the usual manner whenever
necessary. The compositions of the various solutions in which
the tumor fragments were immersed are given in table 1.

c. Growth of transplanted tumors after different treatments

The complete results of the transplantation experiments are
recorded in 39 charts. As it is manifestly impracticable, even
if desirable, to reproduce this number of charts, the results will
be given in the first instance in table 1 as far as possible. Al-
though this method of presentation leaves much to be desired,
especially in the recording of the detailed growths of the various
transplants, it will serve to bring out the points which are to be
emphasized. In order to present a partial idea, at any rate, of
the various growths, a number of more or less typical examples
were chosen from the charts and are recorded in figure 1.

7rowi

h o/ Fie*/

Var

Jb/Zncj F%q£ C^cfrc /riomo Q ft&/°
/ous Treatments.

Series

Rot Ma

come.

trs.

7 ObyS

14 dps, 2i atr/s

26 aips

35 oa/s

A (1)

1156

contnt

•

80S)

1256

S£

buffer

Q(8)

1266

8-2

buffer

•

803)

1267

6.2

buff*

C M

1362

7.0

bufkr

•

CO')

1365

7.0

bof&r

•

C 02)

1310

buf&r

•

00+)

1455

1.0

NaCl

•

D 05)

1462

7.0

L/C/

•

1524

7.0

CoCk

0.5

•

£ 09)

1530

1.0

CoCh

•

FC2')

1551

7.0

CoCfs

F tf$

1560

1.0

CaCfj

•

Ftf3

1562

7.0

CoCfi

£>(*)

1465

7.0

CbC/i

to)
/ (3")

1116
1082

7.0

CoCb

—

•

Nad
CoCh

•

J (35)

1032

NoCl
KCI
CaCk

2044

NoC/

h-HrHri

cm.

Fig. 1. Growth of transplanted Flexner-Jobling rat carcinoma after immer-
sion in various salt solutions and in solutions of different hydrogen ion
concentrations.

* Concentration of salts as in Locke-Ringer solution.

t Concentrations of salts three times those in Locke-Ringer solution.

287

288

K. SUGIURA, H. M. NOTES AND K. G. FALK

i |

^ ft

c3 0) >>

J£ o 5

O H3

^ O

c
u

O
tJO

^ 03

£ o
I o

O _

^ o

C3 £

o o
bfi CO

« 8

a i _

P 5 S

« o

.3 o a.

O o

* -

CM O

_L °

■* O

Sfi. O
to

o
o

<=> g

8 ©

+ 8
_ o

G *

^ d
to O

■s

10 o

+ s

—

s«

. to

o —

+ <N
_ O

O *

£■< d

to o
to

o 9

O ICQ

^ Icq
P5 ©

v Icq
©

pt| ICS jCS

8I« Sh

Ph o K h

PhIcs OS

f=4 I01

|os

-s o i 3 9
5 ,s. 5 J i

BgSSS

GROWTH OF FLEXNER-JOBLING RAT CARCINOMA

o o

° la

02 O -23

« c 1

fa § 2

03 £ bC

03 03

03 ^-J
CP 0)

O 03
*=! &
ft

§ 8

£ O 03

03 03 03

F- 73

03 -C
F-, *a +»

o3 y=!

o
*oa

ft
o
*Q3
>

* &

03 c3
S3 M

03
03

•+3

-t-3

,£

-a

-1-3

«*-c

O o3

bC .-+3

O o3

H-3

Fh
9

-CJ

-1-3

4-3

-+3

Wit

Fh
bC

03 03

a .2

g f

09 C3
^ %

o ^

hS Fh

O «*H

Fh C3

b£) 03

.Jg 03

o8 £
* c3
O

"*> Fh
O

ci
Fh

(3 03

(09)

o
o

o
o
1—1

o
o

1— 1

o
o

to o

03 .

<N g

^ d

jo o

o 43

^ d

w ^

io O

o °

• o

CO §
+ (N

* o
O ^
^ d

iO Q

a is

SI2

fa 1°

fa 1®

»o

— ^

T— (

290

K. SUGIURA, H. M. NOYES AND K. G. FALK

t-l

o
■-

bfl

o
■-

to
"S.

S
O

o
a
O

o o

t-l

be be

5 2

£ * £

W) fl 13
O ^

§ 15

o3 03

o e.
H 5

O ^
Mm

o R

P=H I05

III Sis

gig

Sfa 2 la

£ |cb

> O i 3 Q

55 9 * a p

GROWTH OF FLEXNER-JOBLING RAT CARCINOMA

291

55 oJ

£ B
2 e

id I

O
s_
fcfl

o
6

CO O

d »

o »t

2 °

— <u

£ a

o ^3

£3 ^

-3

C c3

^ g

s §
o

o
•-

fcfi

o
t-l
bfi

5H
<N

£ w o

s s s

CO CO

io o o

T-H O O

o n ^

£ W o
£££

ICS |OJ tn |Ci U,

3l«

Ph ©

is, !©

8I«
Ph ©

Sis

292

SUGIURA, H. M. NOYES AND K. G.

FALK

O c

to G

TO f-c

o o

fcfi H

2 *

13 o

tJD
O

o
u
to
o

o
o

5 1

o
O

bC
O

Sz;

bfl

Sz;

° H 1

e H J

.§"3 6.2

O OS

o o o

O O

8 03
P P

2l«

O ipq

Sis

O itt

Sis

Op op Op
EIS SIS pJS

>i O ■ "i a

5 < a p

GROWTH OF FLEXNER-J OBLING RAT CARCINOMA 293

The data given in table 1 are perhaps sufficiently self-explana-
tory. The experiments were allowed to proceed six to seven
weeks in every case. In the column showing percentage of
positive inoculations, whenever the growth was delayed for a
period greater than one week, it was calculated as a positive
inoculation, but the number indicating the percentage was en-
closed in parentheses. Figure 1 shows the relative growths under
the different conditions more clearly.

The results of the various treatments on the growth of the
transplants can be summarized in a comparatively brief manner
as follows:

1. After immersion in potassium phosphate buffer mixtures of
various hydrogen ion concentration for twenty-four hours, the
following results were obtained: No effect was observable at pH
7.0, the growths being practically the same as in the control; no
growths at all were obtained at pH 5.8 and 5.1; at pH 8.2 partial
inhibition and delayed growths were found; and no growths at
pH 8.8.

2. Varying the concentration of the buffer mixture (from
approximately 0.6 per cent to 1.1 per cent) at pH 7.0 (series
CIO, Cll, and CI 2) and immersing for twenty-four hours did
not produce any effect, the growths being normal in all the cases.

3. Sodium chloride (0.15 M or 0.9 per cent) at pH 7.0 had
no effect in twenty-four hours on the growth of the transplants;
seventy- two hours immersion resulted in complete inhibition of
the growths; lithium chloride at pH 7.0 in twenty-four hours
caused partial inhibition.

4. Calcium chloride at pH 7.0 inhibited or retarded the growths
of the transplants. Immersion for one-half hour had no effect;
for five and ten hours, partial inhibition or retardation; for
twenty-four hours and longer, complete inhibition.

5. Immersion in a Locke-Ringer solution for seventy- two
hours did not affect the growths. Immersion in a solution
containing three times the concentration of the sodium, calcium,
and potassium chlorides in the Locke-Ringer solution for twenty-
four or seventy-two hours resulted in complete inhibition of the
tumor transplants.

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 4

294 K. SUGIURA, H. M. NOYES AND K. G. FALK

d. Histological examinations of tumors

The histological examinations of a number of the tumor trans-
plants after immersion in the various solutions were made by
Dr. James Ewing, to whom the writers wish to express their
thanks. The general structure after immersion in the phosphate
mixtures of various hydrogen ion concentrations showed slight
hydropic degeneration in the case of pH 7.0, but with the more
acid and more alkaline solutions the tumor cells had undergone
marked hydropic degeneration, being larger and considerably
swollen. The microscopical structure of this tumor is shown
in plates 1 and 2; in plate 1 an untreated section, and in plate 2
one treated with a solution of pH 5.8. The histological com-
parison of untreated tumor and tissue which had been immersed
in solutions of sodium chloride, lithium chloride, and calcium
chloride also showed some interesting facts. In those treated
with sodium or lithium chloride, there was considerable accumu-
lation of fluid between the tumor cells and many nuclei appeared
slightly shrunken. On the other hand, tissue which had been
treated with calcium chloride showed imbibition of fluid in the
intercellular spaces and the cells themselves were shrunken.
The nuclei stained more densely and the chromatin was condensed
to a central mass within the nuclei. The cytoplasm showed
hydropic vacuolization (plate 3). A microscopical examination
of tumor tissues was made after immersion for seventy-two hours
in a Locke-Ringer solution, and also of the same tumor tissue
treated by a modified Locke-Ringer solution in which the salt
content was increased three times. Under both conditions the
tissues showed marked hydropic degeneration, but the proliferat-
ing capacity of the tissue treated with normal Locke-Ringer
solution was not altered; while under the second condition
proliferation was completely inhibited.

3. DISCUSSION

In comparing the results of the tumor transplanting experi-
ments, it is necessary to consider not only the contents of the
solutions in which the tumor fragments were immersed before

GROWTH OF FLEXNER-JOBLING RAT CARCINOMA 295

the inoculations, but also the lengths of time of the immersions.
The temperatures were kept fairly constant, otherwise these
also would have to be included. It will be seen that in discussing
the contents of the solution, not only must the salt and its
concentration be considered, but that the hydrogen ion concen-
tration is of the greatest significance.

The favorable effect of immersion for twenty-four hours in
solutions of pH 7.0 as compared with the more acid or more alka-
line solutions is striking. Immersion in the pH 8.0 solution
appeared to be less harmful than immersion in the pH 6.0
solution. Complete inhibition of growth was caused by the
pH 6.0 and pH 9.0 solutions in twenty-four hours. From a
histological examination of the tumor cells after immersion in
solutions more acid or more alkaline than pH 7.0, it was found
that there was a hydropic degeneration of the cells similar
to that of tumor cells resulting from exposure to radium or to
z-rays.

The action of the salts on the tumor tissue before inoculation
is of interest. Sodium chloride had no effect in twenty-four
hours, but exerted a harmful action in seventy-two hours. Lith-
ium chloride produced a small inhibiting action in the twenty-
four hour treatment. Calcium chloride caused a very strong
inhibiting or retarding action on the subsequent growths even
in ten hours. The use of potassium salts in the buffer mixtures
showed that potassium exerted no specific retarding influence.

Cramer (4) several years ago studied the effects of sodium
chloride and calcium chloride solutions on the growth of a trans-
plantable mouse carcinoma. He allowed isosmotic solutions of
these salts to act on the cells for one to two hours before inocula-
tion. Marked inhibition of growth was shown by the cells
immersed in the calcium chloride solution, practically none by
those immersed in the sodium chloride solution. He found that
the harmful effects of the calcium chloride could be overcome
by subsequent immersion in sodium chloride solution, and that
the action of the calcium chloride was only transient.

Although the experiments of Cramer are of interest in con-
nection with the present work in showing a parallelism between

296 K. SUGIURA, H. M. NOYES AND K. G. FALK

the harmful action of the calcium salt and the less harmful action
of the sodium salt, the following differences in the manipulations,
aside from the use of mice in the one case and rats in the other,
may be pointed out. In the present work, the hydrogen ion
concentrations of the solutions in which the tumor fragments
were immersed were controlled; in Cramer's work they do not
appear to have been controlled. In the former, the inoculated
animals were observed for six to seven weeks after inoculation;
in the latter for two weeks. In the former, also, the times of
immersion were varied for a number of the treatments, and cer-
tain salt mixtures were used. In spite of these differences, the
results of Cramer offer valuable contributory evidence in connec-
tion with the transplantation phenomena.

The results of the transplantation experiments described in
this paper are evidently due to cell destruction of some form.
In view of the harmful action of the salts separately, and the
harmlessness of the " balanced" mixture, the action on the cell
membrane or wall appears to be the dominating phenomenon.
Destruction of the membrane or cell wall, or perhaps better,
modification of its permeability, destroys the reproductive
power of the cell. The actions of small amounts of acids and
of bases, and of various salts and mixtures of these, have been
described in connection with other cell structures a number of
times. In the present work, the effects of solutions of different
hydrogen ion concentrations have been made more definite, and
the relative action of sodium and calcium salts at a definite hy-
drogen ion concentration brought out clearly. The explanation
of these actions is, however, the same as that developed in other
connections by a number of workers.

The work on the protease actions of malignant human and
rat tumor extracts presented elsewhere (1) shows some interest-
ing similarities and differences as compared with the transplanta-
tion phenomena presented here. While it is true that the causes
underlying the two sets of phenomena are probably of entirely
different nature, the protease results will be outlined briefly in
comparison with the transplantation results.

GROWTH OF FLEXNER-JOBLING RAT CARCINOMA 297

The optimum hydrogen ion concentration for protease action
and the favorable medium in which tumor fragments on immer-
sion retain their ability to grow when transplanted, correspond
very closely to that represented by pH 7.0, unfavorable condi-
tions for both being reached more rapidly on the acid side than
on the alkaline side. Chlorides of the alkalies in fairly dilute
solution and within short time limits, did not affect appreciably
the protease action or the ability of tumors to grow when trans-
planted, but calcium salts under similar conditions exerted a
very marked inhibiting or retarding action on both. On the
other hand, a definite difference is observed with a mixture of
sodium chloride, calcium chloride, and potassium chloride in
the concentration found in Locke-Ringer's solution. Retarda-
tion of the protease action was caused to the same extent that
the salts separately would cause such action, while with the
transplantations, the salts exerted an antagonistic action toward
each other in the sense that no inhibition of the tumor growth
was observable in the subsequent inoculations.

The two sets of actions, therefore, can be ascribed to different
causes; the transplantation results in the first instance to salt
actions on the permeability of the cell membranes, the enzyme
results to chemical actions of unknown nature on the enzyme
molecule or enzyme grouping of some molecule. The compara-
tive actions of the sodium and calcium salts parallel each other
when used alone, but not when used in mixtures in certain pro-
portions and concentrations. The parallelism, even if due to
different causes, as is probable, is of interest.

Certain facts may be mentioned in this connection, even if
they have no direct connection with the phenomena described
in this paper. In reviewing past work on the inorganic constit-
uents of neoplastic tissue, it may be stated that the calcium
content of rapidly growing tumors was found to be small. Cal-
culating the potassium-calcium ratio of such tumors, large
concentrations of potassium as compared with the calcium were
found, while in old, necrotic, or slowly growing tumors, the con-
centration of calcium was considerably greater in comparison
with that of the potassium (5).

298 K. SUGIURA, H. M. NOTES AND K. G. FALK

Finally, it may be stated that while a definite connection
may exist between the conditions affecting the permeability
of cell membranes, the factors which influence the activities of
intracellular enzymes, and the relations of the inorganic con-
stituents of neoplastic or other tissue, such a connection is still
obscure and can only be hinted at until more definite experi-
mental evidence is available.

4. SUMMARY

The growth of the Flexner-Jobling rat carcinoma was investi-
gated after grafts had been immersed in solutions of various
salts of different hydrogen ion concentration. Calcium strongly
inhibits growth, and a pH of 6.0 appears to be more harmful
than pH 8.0.

The growth of the tumors was compared with the protease
action of extracts of malignant human and rat tumors, and simi-
larities and differences in these phenomena indicated, as well as
possible explanations for them.

REFERENCES

(1) J. Biol. Chem., 1922, liii, 75.

(2) Sugiura, K., and Bexedict, S. R. : Jour. Cancer Res., 1920, v, 373.

(3) Clark, W. M. : The Determination of Hydrogen Ions, 1920.

(4) Cramer, W.: Biochem. Jour. 1918, xii, 210.

(5) Beebe, S. P.: Am. Jour. Physiol.. 1904, xii, 1G7.

Clowes, G. H. A., and Frisbie, W. S. : Amer. Jour. Physiol., 1905, xiv, 173.
Waterman, N. : Arch. Neerlnd, Physiol., 1921, v, 305.

PLATE 1

The plates show three photomicrographs taken with the same magnification.
Plate 1 shows untreated tumor tissue of the Flexner-Jobling rat carcinoma.

GROWTH OF FLEXXER-JOBLING RAT CARCINOMA

K. SUGIUBA, H. M. XOYES AND K. G. FALK

PLATE 1

299

PLATE 2

Plate 2 represents the same tumor after immersion for twenty-four hours in
the phosphate mixture, having the pH 5.8.

;*()()

GROWTH OF FLEXXER-JOBLIXG RAT CARCINOMA

K. SrGIURA, H. M. NOTES AND K. G. FALK

PLATE 2

301

PLATE 3

Plate 3 shows the results after immersion of the same tumor for twenty-four
hours in the 0.078 M calcium chloride solution, having the pH 7.0. Microscopical
examination of the treated tissues showed considerable degenerative changes.
(For further reference see text.)

302

GROWTH OF FLEXNER-JOBLIXG RAT CARCINOMA

K. SUGIURA, H. M. NOTES AND K. G. FALK

PLATE 3

303

PRIMARY SPONTANEOUS TUMORS IN THE KIDNEY
AND ADRENAL OF MICE

STUDIES ON THE INCIDENCE AND INHERITABILITY OF
SPONTANEOUS TUMORS IN MICE

Seventeenth Communication

MAUD SLYE, HARRIET F. HOLMES, and H. GIDEON WELLS

From the Otho S. A. Sprague Memorial Institute and the Department of Pathology
of the University of Chicago

Received for publication May 8, 1922

Primary tumors of the kidney occur not infrequently through-
out the animal kingdom, and, in general, seem to exhibit the
variations and peculiarities seen in human renal tumors. As an
indication of the comparative pathology of renal tumors the
following review of the literature is presented:

Mouse: Few cases of primary renal tumors have been de-
scribed among the great numbers of other tumors found in this
species. Tyzzer (1) found 4 renal growths, which he interprets
as hypernephromas, among 83 primary tumors in mice. One
of these mice, an old female, had also an adenoma of the lung,
a carcinoma of the ovary, and a lymphoma infiltrating both
kidneys; another had an adenoma in the lung. At the time
this paper was written the interpretation of hypernephroma was
more liberal than at present, and the illustrations might now be
interpreted by many pathologists as papillary adenoma of the
kidney, which was Tyzzer's original diagnosis in two of his cases.

Among the 300 mice with spontaneous tumors described by
Haaland (2), there were but two renal tumors. One was a large
growth in the kidney of a twenty-two months old male mouse,
and had invaded a vein but produced no metastases; micro-
scopically the structure was that of an adenocarcinoma. The
second tumor, in a mouse of the same sex and age, resembled

305

306 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

perfectly in structure a human hypernephroma; there were no
metastases, although the lung contained an adenoma. Inocula-
tion of the hypernephroma into 40 mice was without result.

Murray (3) also described a spindle-cell growth surrounding
the kidney without infiltrating it, apparently a primary retro-
peritoneal sarcoma and related to the kidney by position only.

Further than the above we can find no reports of primary renal
tumors in mice, although Stumpf (4) has contributed a discussion
of the behavior of carcinomas inoculated into the kidney.

Rats: Of 103 tumors found in 100,000 rats autopsied in plague
work by McCoy (5), 11 were in the kidney, and were classified as
4 adenomas, 6 carcinomas, and 1 papilloma. Woolley and Wherry
(6) in 23,000 rats found 22 tumors, of which 3 were in the kidney;
all were of renal-cell type. Bullock and Rohdenburg (7), in a
compilation of the literature on rat tumors, found 123 (including
those cited above) of which 7 were fibroadenomas, 1 a papilloma,
and 8 carcinomas of the kidney. Among 32 cases of their own
in laboratory white rats> there were 10 adenomas of the kidney.
Therefore, of 155 rat tumors 26 were of renal origin. A further
case of carcinoma of the kidney in a wild rat was reported by
Beatti (8). Loewenstein (9) described tumor-like papillary out-
growths arising in the pelvis of the kidney, as well as in the
ureter and bladder, of rats infected with Trichodes crassicauda
specifica.

Since sarcomas are much more common than carcinomas in
rats, the fact that all the renal tumors in rats are of epithelial
structure is of interest.

Squirrels: Tn 250,000 ground squirrels (Citellus beecheyi) au-
topsied in plague work, McCoy (10) found 8 tumors, of which
one was described as an angiosarcoma of the kidney. A case of
hypernephroma in the kidney of a grey squirrel (Sciurus caro-
linensis pennsylvanicus) was reported by Fox (11).

Birds: Fowls are rarely subject to renal tumors. Of 880
examined by Curtis (12), 79 had tumors, 5 of which were in the
kidneys; but these were not microscopically corroborated and
there is reason to doubt their true neoplastic nature. In 34
cases of tumor in fowls compiled by Wernicke (13), not one was

TUMORS IN KIDNEY AND ADRENAL OF MICE 307

primary in the kidney although in 3 cases renal metastases were
found. In 852 autopsied fowls Burger (14) found 12 tumors of
which none was in the kidney, although a sarcoma of the ovary
had produced metastases in the kidney in one case.

The review of the literature on tumors in fowls and birds by
Joest and Ernesti (15) who collected 112 cases and added about
50 more, reports no further renal tumors.

Other birds seem to have renal tumors perhaps more frequently,
as is indicated by the following reports: White (16) described a
fibrosarcoma in the kidney of a goose. Fox (17) reported in the
kidney of a male chestnut eared finch (Amadina castanotis) a med-
ullary carcinoma, becoming scirrhous in places-, with metastases
to the lungs> and three cases of benign adenoma of the kidneys in
undulated grass parrakeets (Melopsittacus undulatus). He had
previously reported (18) two cases of papillary cystadenoma and
also a "cyst adeno-carcinoma papilliferum" (19) in the same
species; besides an adenocarcinoma of the kidney in a saffron
finch (Sycalis flaveola), (20) and a spindle-cell sarcoma of the
left kidney with metastasis into the left tibia in a scaly ground
dove (Scardapella squamosa) (21). Seligmann (22) reported
as a diffuse carcinoma a growth which involved both kidneys,
with metastases in the liver and mesentery, in a Chilian pintail
(Dafila spinicauda) which was twenty-six years old. Baird (23)
reported a case of keratinizing epithelioma in the kidney of a
fowl, and Borrel and Masson (24) also described a renal tumor
in a fowl which showed both cylindrical and squamous elements.

Domesticated mammals. Rabbits, which are less subject to
tumor than most mammals, seem to have a relatively large
proportion in the kidney, for Scott (25) stated that of 39 new
growths reported, 5 were benign renal adenomas, resembling
Wilms' tumors in structure.

Swine also seem to be particularly subject to renal tumors, for
of the 12 cases of tumors in swine collected by Sticker, 7 arose in
the kidney, and other cases reported since then indicate the same
tendency. Many of these tumors were of the mixed embryonal
tumor type.

308 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

In horses renal tumors are far from rare. In Sticker's (26)
compilation, of 509 cases of malignant tumors in horses 37 were
in the kidneys. McFadyean (27), in 63 cases of tumor in ani-
mals included 5 in the horse kidney, none with metastases. Of
142 equine neoplasms among 77,224 horses slaughtered in Japan,
there were 9 in the kidney, as compared with 49 in the testicle
(Kimura) (28).

According to Trotter's (29) figures, renal tumors are much
less frequent in cattle, for of 305 cases but 1, a colloid cancer,
arose in the kidney; but Sticker's figures show 10 of 78 bovine
tumors in the kidneys. Steinke (30) described as hypernephroma
a tumor of the kidney in a cow. Roussy and Wolf (31) in their
review on cancer in animals, gave a picture of a tumor from a
bovine kidney resembling a hypernephroma, and stated that
cancer of the kidney in horses resembles that in man, while in
swine the adenosarcoma of the kidney is among the most common
tumors. Cadiot (32) quoted a case of an enormous cancer of the
kidney in a mare.

No reports of cases of renal tumors in sheep can be found.

Of 766 primary cancers in dogs in Sticker's tables 19 were in
the kidney. McFadyean also described two renal tumors in
dogs, each case exhibiting metastasis. No renal tumors appeared
in the 21 cats with tumors in Sticker's lists, nor in Roffo's 7 cases
(33) nor in Murray's 11 cases (34) but Teutschlaender mentioned
the cat (Kater) on his list of animals in which carcinoma of the
kidney has been reported. Murray also described 48 cases of
tumors in dogs, 12 in horses, and 18 in cows, with none arising in
the kidneys.

We are indebted to Dr. L. E. Day for his summary of 316
tumors found among 2000 animal specimens sent to the Chicago
Laboratory of the Bureau of Animal Industry. These do not
represent all the tumors that are observed in the slaughter house,
but merely specimens sent to the laboratory by the inspectors
when in doubt concerning the diagnosis. Among 175 tumors of
cattle there were no renal tumors although there were four adrenal
tumors, two diagnosed as sarcomas and two as hypernephromas.
Among 93 tumors in pigs no less than 52 were in the kidney,

TUMORS IN KIDNEY AND ADRENAL OF MICE

309

47 being diagnosed as embryonal adenosarcomas and 5 as sar-
comas ; there were no adrenal tumors. In considering the tumors
in swine it is to remembered that nearly all swine are
slaughtered before they are two years old, so that not many of
the slaughtered animals have reached an age for developing
carcinomas. Among 48 tumors from sheep none whatever were
found in the kidney or adrenal. These figures emphasize the
infrequency of renal tumors in cattle and sheep and the fre-
quency of mixed tumors of the kidney in swine, which resemble
the typical mixed tumors of the human kidney.

Other incidental cases are: Sarcoma of the kidney in a swine
(Hamburger) (35), the report of which leaves some uncertainty
whether the growth may not have been leukemic.

Papillary adenoma in the left kidney in a mongoos lemur
(White) (36). A tumor involving the kidneys, liver, lungs, and
ovaries of a python, the primary site not being determined
(Bland-Sutton) (37). A leiomyoma in the kidney of a horse,
described by Wells (38). Pick (39) described a soft adenoma
involving both kidneys of an eel. Williams (40) stated "A
malignant renal tumor (derived from an adrenal 'rest') has been
described by Bland-Sutton in a marmot." In Teutschlaender's
compilation (44) were listed the following cases from the literature :
Hemangioma of the kidney in a horse; carcinoma of the kidney
in fowl, frog, pike, and buzzard.

While the above compilation is certainly not entirely complete,
it serves to bring together much of the literature on the com-
parative pathology of renal tumors, and to indicate their general
distribution and frequency.

Adrenal tumors. We can find no reports of primary adrenal
tumors in mice, despite the great numbers of mice that have been
examined at autopsy and the not infrequent occurrence of ad-
renal tumors in other species. Thus, Kimura (42) found in the
literature records of adrenal tumors in 24 horses and 46 cattle,
and he himself found 5 adrenal tumors among 46 tumors from
horses. The extensive discussion of the comparative pathology
of adrenal tumors by Steinke (43) mentioned tumors only in
horses and cows, commenting on the absence of such tumors in

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 4

310 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

other species except for an adrenal tumor of parasympathetic-cell
character in a two-year-old swine, reported by Klawitter.

There is, however, the interesting case reported by Smallwood
(44) as a carcinoma of the kidney of a frog and believed to be
derived from the adrenal tissues. Murray, who reexamined this
specimen, corroborated its carcinomatous nature, but his state-
ment suggests that he was not convinced that it developed from
adrenal rather than renal cells. The illustrations in Smallwood' s
article suggest a papillary renal tumor.

Fox (45) described as a hypernephroma a tumor that arose in
the adrenal of a California hair seal (Zalophus calif ornianus) ,
and also described a hypernephroma in the adrenal of a brown
cebus (Cebus fatuellus) (£6).

As stated previously, in the 316 animal tumors examined by
Day in the Chicago Stock Yards, but 4 arose in the adrenal, all
in cattle, 2 being diagnosed as sarcoma and 2 as adenoma.

RENAL TUMORS IN MICE IN THE SLYE STOCK

In 33,000 autopsies performed on mice of the Slye stock, but
16 cases of unquestionable primary tumor arising in the kidney
have been observed, supporting the impression that the kidney of
the mouse is not among the common sites of primary neoplasm.
These renal tumors are, according to the histological evidence, to
be classified as follows: 1 carcinoma, 3 adenomas, 1 hyperne-
phroma, 7 sarcomas, 3 mesotheliomas, and 1 sarcoma in the
renal pelvis. It will be noted that we have observed no in-
stances of the mixed tumors of the kidney that are so common
in man and some other species, nor have we had any papillary
tumors of the renal pelvis. The features presented by these
several types of renal tumor are best described by giving briefly
the findings in each of the several cases that we have observed.

EPITHELIAL TUMORS OF THE KIDNEY

1. Primary carcinoma of the kidney (fig. 1). Female mouse
(1934). This mouse showed a white solid mass in the right kidney
about the size of a normal kidney. Beyond this there were no

TUMORS IN KIDNEY AND ADRENAL OF MICE

311

abnormalities found, except that the spleen was about three times
the normal size. The left kidney was normal, and no metastases
were found. Microscopically the tumor is composed of cells of
epithelial type, but without the characteristics of either adrenal
cells or of tubular epithelium of the kidney. They are arranged

Fig. 1. Primary Carcinoma of the Kidney

The junction of the compressed renal tissue and the tumor is shown. Appar-
ently this tumor is derived from renal epithelium. Mouse 1934. X 110.

in large, pseudo-alveoli with much central necrosis, so that the
persistence of the tumor cells near the bands of stroma gives in
places the impression of a papilloma. Although the growth
does not infiltrate the kidney very much, it has no capsule of its
own, infiltrates the renal capsule in places, and is undoubtedly

312 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

malignant. Evidently it is a tumor derived from the renal
epithelium and may be properly designated as a carcinoma,
although the term mesothelioma might also be appropriately
used. Presumably it is in an early stage of malignancy in view
of the relatively small amount of extension of the growth.

2. Solid adenoma arising in bilateral cystic kidneys. Male
mouse (^9907), age two years, thirteen days. Both kidneys were
converted into masses of small cysts of various sizes, resembling
the congenital cystic kidneys, although no cysts were found in
the liver. Some of the cysts contain colloid masses, and between
them are occasional foci of small round cells. There remains
more kidney tissue in a functional condition than is usually
seen in fatal human cases of congenital cystic kidney, but this
tissue is far from normal, the tubules containing many hyaline
casts, the interstitial tissue being infiltrated with round cells,
and many of the glomeruli being more or less hyalinized. In the
lower pole of each kidney was a fleshy nodule about 3 to 4 mm.
in diameter. Microscopically these nodules are composed of a
solid tissue, made up of groups of large epithelial cells somewhat
resembling renal epithelium. About these is a delicate stroma
containing some collections of small round cells. The nodule is
distinctly encapsulated and seems to be in the renal substance
rather than within one of the cysts. In many respects the
structure resembles that of the benign ovarian adenomas of
mice (47). No abnormalities of importance were noted in the
other organs.

3. Adenoma of the kidney (fig. 2). Male mouse (10220), which
died with advanced sarcosporidiosis ; had marked chronic nephri-
tis, both kidneys being large and nodular, with atrophic areas
alternating with areas of swollen and dilated tubules, some of
the latter being distinctly cystic. Beneath the capsule of one
kidney was a nodule 4 mm. in diameter, with a well defined
capsule. It consists of a mass of large epithelial cells with solid
cytoplasm, arranged in large cords with a very delicate stroma
between them. The arrangement of the cells resembles that of
the adrenal adenomas, and a diagnosis of benign hypernephroma
might readily be made. The cells are less vacuolated than those

TUMORS IN KIDNEY AND ADRENAL OF MICE

313

of the typical adrenal cortex, having a denser cytoplasm, and are
not dissimilar to the epithelial cells found lining the distended
renal tubules; it seems probable, therefore, that this adenoma is
derived from renal epithelium. There are some clefts containing

Fig. 2. Adenoma of Kidney

This somewhat resembles a tumor derived from adrenal cells, but is believed
to have been derived from renal epithelium. Mouse 10220. X 65.

colloid material, and numerous calcific granules in the form of
calcospherites.

4. Adenoma of the kidney. Male mouse (24073), with a
large liver cyst containing a tapeworm, exhibiting in one kidney
a solid subcapsular nodule 5 mm. in diameter, with a delicate but
definite capsule. This is composed of a solid mass of large

314 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

epithelial cells with abundant foamy cytoplasm and small dark
nuclei. The cells are in large cords or alveoli with a very deli-
cate stroma between them. There are numerous small foci of
calcification of the tumor tissue. The foamy character of the
tumor cells, together with their arrangement, suggests an adrenal

Fig. 3. Renal Hypernephroma

This growth entirely replaced one kidney, but produced no metastases; it
corresponds in structure to human hypernephromas. Mouse 3639. X 110.

origin for this growth, but the post-mortem changes have so
altered the auclear and cytoplasmic details that this cannoi be
determined positively. In places then4 is a tendency to tubular
arrangement of the tumor cells, suggesting a renal origin. Out-
side the tumor the kidney shows an advanced chronic nephritis
with many hyaline and fibroid glomeruli.

TUMORS IN KIDNEY AND ADRENAL OF MICE 315

5. Hypernephroma in the kidney (fig. 3). Male mouse (3639)
with no other lesions of significance. The left kidney was re-
placed by an encapsulated, hemorrhagic mass, measuring 22 x 22
x 18 mm. The right kidney was slightly enlarged and soft,
and showed a moderate degree of chronic nephritis. No remains
of renal tissue are found in the tumor mass, which presents the
typical appearance of a hypernephroma. More than half the
mass contains no cells, being composed of the residue of old
hemorrhages and necrosis. The living portions consist of large
foamy cells arranged in cords and alveoli with a delicate stroma.
It is completely surrounded by a capsule which shows no inva-
sion by tumor cells. In all respects this growth corresponds
perfectly to the human renal hypernephromas. No metastases
can be found.

In the hilum of the right kidney, attached to a large artery, is
a nodule about 2 mm. in diameter which, in the center, resembles
a small leiomyoma, but about it is a mass of granulation tissue.
It bears no resemblance to the hypernephroma and its nature is
unknown.

In view of the fact that chronic nephritis is one of the com-
monest diseases in mice, and appears in forms often quite similar
to nephritis in man, it is strange that we have found so few in-
stances of epithelial neoplasms in mice, especially the benign
adenomas which are so often found in human kidneys showing
chronic nephritis.

SARCOMA OF THE KIDNEY

The diagnosis of sarcoma, always difficult and often unsafe,
is particularly dubious in the case of the kidney which presents
so many non-sarcomatous growths that resemble sarcoma, and
especially in the face of the statement made by Ewing (48) :
"Birch-Hirschf eld's group of adeno-angiosarcoma, derived from
Wolffian remnants, and the lipomyosarcomas remain, however,
the only well-defined varieties of renal sarcoma which have been
fully divorced from a probable epithelial origin." Nevertheless,
since the kidney contains connective tissue it is perfectly possible
for sarcomas to arise therein, and in this mouse material we have

316 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

several specimens for which only the diagnosis of sarcoma can
be made, after excluding all other possibilities. In doing this
we have carefully eliminated numerous growths of doubtful
character, some of which also may really be sarcomas.1 These
cases are briefly described as follows:

6. Bilateral sarcoma of kidneys. Male mouse (7667), with
both kidneys symmetrically enlarged to equal size, about 17 x 10
x 9 mm. They contained little recognizable kidney tissue, and
were for the most part infiltrated by a fleshy, pinkish white
tissue which involved equally the cortical and medullary portions.
No lesions were found elsewhere. Microscopically both kidneys
show infiltration replacing about 80 to 90 per cent of the renal
elements. The neoplasm is composed everywhere of slightly
oval cells, a little larger than lymphocytes, with deeply staining
nuclei and very little cytoplasm. No evidence of neoplastic
epithelial or mixed tumor elements can be found. This tissue
infiltrates between the tubules much as do the cells in leukemic
infiltrations, and the capsule is also invaded. The diffuseness
of this infiltrative growth and its equal involvement of both
kidneys makes it resemble a leukemia or pseudoleukemia, but
this diagnosis is untenable in view of the lack of involvement of
other organs or lymph-nodes. The diagnosis of sarcoma is made
largely by exclusion.

1 We wish to quote here a statement of the criteria used in our consideration
of sarcoma throughout this work as expressed in our paper on Primary Spontane-
ous Sarcoma in Mice (Jour. Cancer Res., 1917, ii, 1). "We recognize fully the
difficulties that attend the differentiation of sarcoma, and for the purpose of
this study have excluded every form of new growth concerning the nature of
which there seemed any possible room for question. Therefore, we have not
included numerous cases in which we think that the growths are probably
sarcomatous, and many more in which we cannot be sure that the neoplasm
is not sarcoma. On the other hand, the statistical value of our figures is
lessened by the fact that we have undoubtedly omitted some growths that are
t rue sarcomas. Our figures represent minimal values only. From the standpoint
of investigations in heredity, with which our work is particularly concerned,
it is just as undesirable to call a sarcoma something else as to include a granuloma
among the sarcomas, and hence the rigid classification adopted in this study of
sarcomas is no more satisfactory for our heredity statistics than would be a lax
classified ion that included some growths of doubtful nature. Therefore, in
charting the heredity statistics it is necessary to recognise the absence of positive
criteria for the differentiation of sarcoma, and to admit the borderline cases with
a mark of interrogation to indicate this fallibility."

TUMORS IN KIDNEY AND ADRENAL OF MICE 317

7. Sarcoma of kidney. Old male mouse (24979), with the
right kidney entirely replaced by a tumor, 16 x 12 x 10 mm.,
which infiltrated the adjacent tissues, including the ureter and
the pelvis of the opposite kidney. The regional lymph-nodes
were also invaded, and there was a mass in the root of the mesen-
tery about 20 mm. in diameter, which seemed also to infiltrate
the pancreas.

Microscopically, the right kidney is found to be almost com-
pletely replaced by a growth of large round cells, uniform in
size, with very little cytoplasm, but with nuclei much larger
than those of lymphocytes. There is no tendency to structural
arrangement, the tumor cells infiltrating freely the tissues about
the kidney, including the adjacent muscles and the pelvis of the
opposite kidney, the ureter of which is surrounded by a mass of
tumor. The regional lymph-nodes are replaced by tissue of the
same character, but the mesenteric mass is completely necrotic,
with few cells resembling those of the tumor. By virtue of its
highly infiltrative character and the large size of the cells, this
seems to be an undoubted sarcoma. The other tissues showed
no evidence of either leukemia or pseudoleukemia.

8. Sarcoma of the kidney. Female mouse (26867) with much
sarcosporidiosis, showed a marked enlargement of the left kidney,
which was between two and three times the normal size, and
infiltrated diffusely with a pink, fleshy tissue. Elsewhere in the
body there were no important changes. Microscopically it is
found that the enlarged kidney is infiltrated extensively, nearly
all the renal elements being replaced by a tissue composed of
round and elongated cells, considerably larger than lymphocytes
and with more cytoplasm. They tend to be arranged in wide
bands, growing out from the blood-vessels, but this arrangement
is not constant or well developed. The neoplastic tissue grows
out from the pelvis and tends to invade the adjacent tissues and
the hilum of the opposite kidney. No similar tissue is to be
found elsewhere in the mouse. The probable diagnosis is pri-
mary sarcoma of the kidney.

9. Sarcoma of the kidney. Female mouse (27148), with the
right kidney much enlarged (18 x 15 x 32 mm.) and consisting of

318 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

a firm tissue overlaid with softer portions. Microscopically it is
found that the kidney is diffusely infiltrated with a growth con-
sisting of polyhedral cells, smaller than epithelial cells usually
are, but with a little more cytoplasm than lymphoid cells. These
cells exhibit no structural organization and have replaced all but
a few of the original renal elements. About one-third of the
tissue is necrotic. A small amount of the neoplastic tissue
infiltrates the hilum of the left kidney, which also shows amy-
loidosis. No similar tissue is found elsewhere in the body, and
there are no evidences of diffuse lymphoid hyperplasia or similar
conditions. The lung contains a small benign adenoma.

The tumor shows much variation in the size of the nuclei,
hyperchromatism being common, and occasional very large cells
are seen. Unfortunately there is too much post-mortem change
for exact details to be seen. The general character of the growth
and the absence of similar changes in other tissues make a diag-
nosis of sarcoma seem inevitable.

10. Bilateral sarcoma of kidneys, with metastasis in spleen.
Female mouse (396), with the uterus enormously distended with
fluid because of vaginal obstruction by a seminal mass, and with
great enlargement of both kidneys by a uniform white tissue.
The right kidney measured 18 x 13 mm., the left 15 x 13 mm.
On the anterior surface of the left kidney is a depressed scar.
In the spleen there is a tumor nodule, 4x5 mm. No other tumor
growths were found. The adrenals were normal.

Microscopically, both kidneys are found to be infiltrated dif-
fusely, with replacement of 80 to 90 per cent of the renal tissue
by a growth uniformly composed of small oval cells in solid
masses, without any particular organization. No similar tissue
can be found in any other organ except the spleen, which has a
tumor nodule of the same structure as the renal tumor. The
adrenals are in contact with, but not invaded by tumor. The
tumor forms large masses in the hilum of one of the kidneys, or
within the kidney itself. The left kidney is more involved than
the right, and probably was the starting point of the sarcoma.

11. Lymphosarcoma of kidney. Female mouse (13124), with
two separate mammary carcinomas, and in the left kidney a

TUMORS IN KIDNEY AND ADRENAL OF MICE 319

nearly spherical mass about 14 mm. in diameter, which was
partly necrotic. Microscopically this growth is composed of a
dense mass of small round cells without visible cytoplasm,
packed closely together without attempt at formation of any
structure. It infiltrates the adjacent renal tissue freely, and there
is a small amount of infiltration of the hilum of the opposite
kidney. The structure is of distinctly neoplastic character, and
there is no similar infiltration of other organs, or lymph-node
enlargement, such as characterizes pseudoleukemia. The diag-
nosis of sarcoma, in structure corresponding to lymphosarcoma,
is the only one that can be made on the features presented.

12. Lymphosarcoma of kidney. This seems to be similar to
case 11, but unfortunately post-mortem decomposition has ad-
vanced so far that accurate microscopic study is not possible.
Such tissue as remains stainable resembles a lymphosarcoma.
The mouse was a female (12533) with no other tumors, and the
right kidney was replaced by a soft pink mass measuring 14 x 12
x 12 mm. A small amount of the same tissue invaded the hilum
of the left kidney.

MESOTHELIOMA OF THE KIDNEY

In our previous papers on tumors in the ovary (49) and testicle
(50) of mice we have called attention to the not infrequent
occurrence of tumors composed of polyhedral cells, presenting
some characteristics resembling carcinoma and some resembling
sarcoma, (figs. 4, 6, 7 and 8) and hence suitably designated meso-
thelioma, in accordance with Adami. These tumors are charac-
teristic of the urogenital anlage, and hence it is not surprising
that growths of the same structure are found in the adrenal and
kidney. We have observed the following cases of renal tumors
that seem to belong to this group.

13. This is a remarkable case in that a female mouse (21663),
when but one month old exhibited two independent carcinomas of
the mammary gland, and osteosarcomatous growths in the
spinal column near the pelvis and in the left fifth rib. It lived
but eighteen days more and at autopsy showed its left kidney
also almost completely replaced by a tumor measuring 18 x 12 x

320 MAUD SL YE , H. F. HOLMES AND H. G. WELLS

10 mm.; the right kidney, similarly involved, measured 10 x 6 x 6
mm. (fig. 4). These renal growths are entirely different in
structure from the other four tumors. They are alike and vary
in appearance in different parts, some portions resembling

Fig. 4. Mesothelioma or Kidney

This portion of the growth presents a sarcomatous character; other portions
exhibit more resemblance to epithelial growth. The tumor was found in a mouse
but one month old, with four other tumors, which if not congenital must have
developed very soon after birth. Mouse 21GG3. X 110.

spindle-cell sarcoma while others are composed of larger, more
polyhedral cells, arranged in a somewhat alveolar fashion, often
separated by highly vascular septa. This structure corresponds
to the type of growth often seen in tumors of the ovaries, testicles,

TUMORS IN KIDNEY AND ADRENAL OF MICE 321

and adrenals in mice, and agrees with the tumors called meso-
thelioma by Adami and Woolley. Since, except for this case,
there have been very few other cases of malignant tumors
arising in mice less than six months of age, the occurrence in so
young an animal of at least four independent primary growths
representing three distinct types of malignant neoplasm, is a
most remarable condition, without, as far as we know, a parallel
in either mouse cancers or in those of any other animal.

14. Male mouse (10011), with no other lesions of importance,
had its right kidney largely replaced by a whitish tumor, forming
a mass measuring 14 x 12 x 10 mm. Microscopically the tumor
has largely replaced the kidney and has infiltrated the capsule in
places. It is composed of small cells with a dark, round nucleus
and a small amount of cytoplasm. These cells tend to form
bands or pseudo-alveoli in some places, but for the most part the
growth is composed of cells in a structureless mass. The left
kidney was not involved by the tumor.

15. This seems to be similar to case 14. Male mouse (9779)
had an enlarged left kidney without other lesions of note. Micro-
scopically the kidney is largely replaced by a mass composed of
small cells slightly larger than lymphocytes and with more
cytoplasm, showing no structural arrangement. The main neo-
plastic mass lies at one side of the kidney, which it infiltrates
slightly. The kidney itself shows some foci of round-cell in-
filtration. Some of these areas slightly resemble the tumor, but
probably are not a part of it. There are some small areas of
calcification, some scars, and numerous hyaline casts. Post-
mortem changes are too advanced for more accurate study.
The opposite kidney contains no tumor, but there is the same
amount of amyloid and calcification.

SARCOMA OF RENAL PELVIS

We have excluded numerous cases in which a retroperitoneal
tumor of sarcomatous character has invaded the kidney hilum,
but there is one case in which both the gross and microscopical
findings distinctly indicate that the tumor had its origin in the
tissues of the pelvis itself.

322 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

Female mouse (348) had a pale swelling extending downward
and inward from the pelvis, composed of tissue of about the same
consistency as the kidney and of uniform yellow color. On cross
section the pelvis of the kidney was in the center of the mass
formed by the tumor and the kidney, which measures 12x9 mm.
The left kidney was of normal size with a pale area of swelling
on the anterior surface. Xo other changes of importance were
found except severe edema of the lungs.

Microscopically the growth is found to lie symmetrically
about the pelvis of the kidney and the upper end of the ureter,
invading the wall of these structures extensively. From here it
passes along the vessels deeply into the kidney, but does not
extend far into the cortex. The kidney is about one half as large
as the tumor, the tubules being much dilated and the glomeruli
more or less hyalinized. The tumor also extends some distance
along the renal capsule as a thin layer. As no growths are found
elsewhere it is evident that this growth arose in the tissues about
the hilum of the right kidney. There is only a small amount of
invasion of the left kidney hilum. The growth is composed of
large round cells with a delicate reticulum. There are several
delicate vessels packed with lymphoid cells, but no other evi-
dences of lymphatic origin in this tumor. The cells of the tumor
are polymorphous with considerable cytoplasm, and deeply
staining nuclei much larger than those of the ordinary lymph-
cell, and usually larger than the nuclei of the renal epithelium.
Xo mitoses are seen. No changes of importance are found in the
other organs.

This tumor is much less complex in structure than the cases of
sarcoma of the renal pelvis in children described by de Vecchi
and Salomon (51).

ADRENAL TUMORS

As mentioned above, no cases of adrenal tumors have hitherto
been described in mice. We are able to report a few unquestion-
able instances of such tumors. Despite the frequency of benign
cortical adenomas of the adrenal in man, we have found but one
such tumor in mice. It may be recalled that adrenal tumors
have not been found frequently in other species of animals.

TUMORS IN KIDNEY AND ADRENAL OF MICE

323

Cortical adenoma of misplaced inter-renal adrenal rest (fig. 5).
Between the kidneys of a female mouse (1921), with marked
amyloidosis and chronic nephritis, was found a spherical mass
about 5 mm. in diameter, attached neither to the kidneys nor to
the intestines. Both adrenals were present at their normal site

Fig. 5. Cortical Adrenal Adenoma

The tumor developed in an inter-renal rest of adrenal cortex tissue. A small
zone of compressed adrenal tissue is seen at one side. Mouse 1921. X225.

and were of normal structure and size, except for some peripheral
round-cell infiltration.

The tumor nodule is composed of a solid mass of cells resemb-
ling those of the adrenal cortex except in lack of orderly arrange-
ment, closely packed together, and flattening out a thin shell of

324 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

adrenal cortex, evidently all that remains of an adrenal rest, since
no medullary elements are to be found. In all respects this
tumor corresponds to the simple adenoma of the adrenal cortex
seen in man.

MESOTHELIOMA OF THE ADRENAL

This seems to be the commonest tumor of the adrenal, as also
of the testicle, of mice, and it is quite impossible to distinguish
on the basis of microscopic appearance between the mesothe-
liomas arising in the different organs derived from the urogenital
anlage. The cases in which the diagnosis seems certain are the
following :

1. Male mouse (10390) had in place of the right adrenal a
spherical mass 5 mm. in diameter, which is completely encapsu-
lated and does not involve the kidney. No other nodules or
findings of importance elsewhere. Microscopically the nodule
contains no remains of adrenal tissue, but consists of a solid tumor
made up of masses of large cells with considerable cytoplasm and
large oval or spherical nuclei, arranged in atypical alveoli or
broad bands with a very small amount of stroma containing
thin-walled blood-vessels. In numerous places the capsule is
infiltrated with tumor cells and there is some invasion of the
areolar tissues about the adrenal, but no invasion of the adjacent
kidney. This seems to be a typical mesothelioma of the adrenal
in an early stage of malignancy.

2. Mesothelioma of adrenal with peritonea1 metastasis. Fe-
male mouse (12744), with abdomen greatly distended by a bloody
exudate, presented at the site of the left adrenal a whitish mass
about the size of the kidney. All through the abdominal cavity
are masses of partially necrotic whitish tissue, especially attached
to the liver and uterus, which do not seem to be involved by this
growth. There are enlarged retroperitoneal lymph-nodes, but
elsewhere no lymphatic involvement. There are no tumors in
the lung's.

Microscopically the tumor shows the usual features of the
typical mesotheliomas, as described in the other cases, with slight
tendency to alveolar arrangement. It does not seem to have

TUMORS IN KIDNEY AND ADRENAL OF MICE 325

infiltrated or produced metastases in the viscera. The right
adrenal is surrounded, but not invaded, by the tumor and seems
normal. The retroperitoneal nodes are replaced entirely by
tumor tissue.

Fig. 6. Mesothelioma of Adrenal

This growth surrounded the lymph nodes and was widely disseminated through
the perirenal tissues. Mouse 7699. X 110.

3. Bilateral malignant mesothelioma of adrenals (fig. 6).
Female mouse (7699) with much subcutaneous edema, showed
numerous enlarged subcutaneous lymph-nodes, some red and
some pale, the largest measuring as much as 8 x 10 mm. There
was a milky fluid in the peritoneal cavity, presumably because
of pressure of enlarged lymph-nodes on the thoracic duct, for

326 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

the retroperitoneal nodes were greatly enlarged, up to 6 x 12 mm.
The mesenteric nodes were also enormously enlarged, one mass
at the root of the mesentery measuring 40 x 20 x 20 mm. Both
adrenals were greatly enlarged, each being about the size of the
kidney, which was adherent to the adrenal but not enlarged or
infiltrated by neoplastic tissue. The uterus seemed to be in-
filtrated by tumor. The liver was not affected, and the lungs
showed only a single small nodule, although there was some tissue
increase in the upper mediastinum; both lungs showed a bloody
edema, and there was a bilateral hydrothorax.

Although the gross appearance suggested a general lympho-
sarcomatosis or pseudoleukemia, yet the microscopic structure
is of an entirely different character. Everywhere the neoplastic
tissue presents the same appearance, being composed of a solid
growth of cells with considerable cytoplasm, so that they look
much like epithelium. The nuclei, which vary greatly in size,
are much more solid than those of epithelial cells usually are.
Mitotic figures are abundant. The invaded lymph-nodes are
largely replaced by tumor cells, which are also found in the
lymph-vessels of the lungs, but not in the liver. Both adrenals
seem to be entirely replaced by the tumor, which lies upon the
capsule of the adjacent kidney without any invasion of this
organ. There are some areas of hemorrhage, but not much
necrosis. The cells show no attempt at definite arrangement,
but simply form a solid mass with numerous, poorly defined blood
spaces.

This tumor resembles in structure other growths found arising
in the urogenital anlage, and in view of this and of the complete
replacement of both adrenals it is most probable that it did arise
in these tissues, although the extensive lymph-node involvement
is unusual in adrenal tumors. The structure is not essentially
dissimilar to that of the adrenal tumor 10360, except for the
amount of extension.

UROGENITAL MESOTHELIOMAS OF UNCERTAIN ORIGIN

Because of the identity in appearance of mesotheliomas from
all organs originating in the urogenital anlage, it is not always

TUMORS IN KIDNEY AND ADRENAL OF MICE

327

possible to decide the place of origin of some tumors which in-
volve two or more of these organs, as shown by the following
cases.

12307. Mesothelioma of either adrenal or ovary. It is not
possible to determine the origin of this tumor, which we have
described in our paper on tumors of the ovary in the following
words :

The abdominal cavity shows several nodules whose exact origin is
difficult to determine as the mate has partly devoured the body. The
right ovary is, however, easily distinguished. It measures 18 x
12 x 12 mm. What seems to be the left ovary is 10 x 8 x 6 mm.
There are 8 other similar nodules in the abdominal cavity, one being
in the position of the left adrenal, measuring 10 x 8 x 8 mm. The
other nodules are apparently in the mesentery. One lobe of the liver
is converted into a tumor nodule 14 x 10 x 18 mm., irregular and
lumpy in outline, pink in color.

The tumor shows everywhere the same structure, consisting of
irregular alveoli composed of large cells with abundant cytoplasm with
well defined borders and deeply staining nuclei. Mitotic figures are
numerous. The character is that usual to mesothelial growths. The
ovary cannot be positively identified, but one mass exhibited in the
capsule structure suggests compressed ovarian tissue with degener-
ated ova. In all respects this tumor is identical with the malignant
ovarian tumors just described.

It seems probable that this tumor arose in the ovary which exhibited
the largest growth, but it is not possible to exclude the adrenal as the
primary site.

The malignant tumors of the adrenal, ovary, and testicle com-
monly exhibit identically the same histological picture as that
seen in this case.

We have also observed two other cases, described in the paper
on ovarian tumors in which we cannot state whether the renal
growth was primary or secondary.

12876. The left kidney contained a mass of pink, fleshy
tissue, 18 x 14 x 14 mm. The right kidney, which was slightly
enlarged, contained no tumor. The right ovary consisted of a
pinkish tissue resembling that in the kidney, and measured

328 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

12 x 8 x 8 mm. In the mesentery was a similar, slightly paler
mass, 16 x 8 x 8 mm. The retroperitoneal and subcutaneous
nodes were not enlarged and no nodules were found in the lungs.

Microscopically the tumor is alike in all three places, consisting
of a diffuse infiltrating growth of large round cells, which also
invade the connective tissues about the kidney and ovary. It
does not at all resemble the typical ovarian tumors, being ap-
parently a round-cell sarcoma. We have no way of telling which
of the three tumors was primary. The next case presents similar
difficulties.

26. This mouse had a tumor mass about 8 x 10 mm. in the
upper portion of the liver, with other smaller nodules near it.
A similar small nodule was found in the right kidney. The right
ovary was enlarged to two-thirds the size of a kidney, and was
solid. Microscopically all these growths are composed of round
cells, apparently a round-cell sarcoma. It is impossible to say
which growth was primary.

Among several tumors arising in the retroperitoneal tissues,
mostly of sarcomatous type, are a few of a structure identical with
the characteristic mesothelioma type of growth that arises in
the urogenital anlage. In the two cases described below this
character of growth was so marked that it seems probable that
the tumors have arisen in some misplaced embryonic rest, since
the organs of this series were distinctly not the starting point of
the growths. Numerous tumors of this sort have been described
in man.

(22380) Malignant retroperitoneal mesothelioma with exten-
sion through the body wall. Male mouse presenting externally
a subcutaneous tumor, involving the left hip and extending
to the anus, the external measurements being 30 x 25 x 25 mm.
W hen the body was opened it was found (hat this mass was an
extension of a huge mass (40 x 20 x IS nun.) which had sur-
rounded the left kidney and pushed it to the ventral midline,
invaded the spleen and pushed it mostly to the right of the mid-
line by formation of a mass that measured 20 x L5 x L3 mm., and

extended downward t brough the pelvic cavity where it surrounded
the rectum and seminal vesicles and broke out through the

TUMORS IN KIDNEY AND ADRENAL OF MICE

329

body wall. There was slight infiltration of the lower pole of the
right kidney. The liver was slightly enlarged and mottled with
areas of infiltration. There was some tumor in the lower part of
the mediastinum, and the lungs show many small spots of neo-
plastic tissue. There seemed to be no involvement of the

Fig. 7. Secondary Mesothelioma in Liver

The primary growth was either in the adrenal or in a retroperitoneal embryonic
rest; it infiltrated the body wall and retroperitoneal tissues, and produced innu-
merable metastases in the lungs and liver. This section shows the masses of tumor
cells compressing and replacing the larger liver cells. Mouse 22380. XI 10.

lymph-nodes, either subcutaneous or abdominal. The testicles
were not involved.

Microscopically the tumor everywhere consists of masses of
large cells with much deeply staining cytoplasm, somewhat
resembling liver cells, with a slight tendency to arrangement in

330 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

cords or bands. Where the tumor invades the liver it can be
seen that the cells are a little smaller and paler than the liver
cells (fig. 7).

The growth infiltrates the liver widely, invades the muscle
of the body wall, invades the renal capsule but not the kidney,

Fig. 8. Secondary Mesothelioma of Lungs

From the same case as figure 7. The section shows the extent of involvement
of the lung. Mouse 22380. X 60.

infiltrates the abdominal sympathetic ganglia, surrounds densely
the rectum and spreads into the seminal vesicle, invades the
spleen, and in the lung appears as multiple tumor cell emboli
within the vessels and as large tumor nodules largely replacing
portions of the lung ffig. 8). The left adrenal was found entirely
embedded in but not infiltrated by the tumor.

TUMORS IN KIDNEY AND ADRENAL OF MICE 331

The microscopic appearances are identical with those usual
in malignant adrenal tumors, but the adrenal was not involved;
and in view of the retroperitoneal origin, the growth may be
presumed to have arisen in an embryonal rest of the urogenital
anlage.

(9979) Malignant retroperitoneal mesothelioma infiltrating kid-
neys. Small female mouse with extensive infiltration of the right
thigh by a typical spindle-cell sarcoma, and with a mass above
the left kidney, about as large as the kidney itself, which was
infiltrated by the growth. The entire mass measured 20 x 14 x 10
mm. The right kidney was not so much enlarged. The growth
seemed to have arisen at the site of the left adrenal, and to have
pushed the kidney forward. The liver was invaded by tumor
and enlarged. No metastases were found elsewhere. Micro-
scopically the tumor about the kidney bears no resemblance to
the spindle-cell sarcoma of the thigh, being composed of a mass
of polyhedral cells arranged in solid masses, with a slight tendency
to form bands or cords. The nuclei are not much larger than
those of lymphocytes, but the cells have much more cytoplasm.
The chief mass lies above and behind the left kidney, which shows
considerable invasion through the capsule and about the blood
vessels. There is a similar invasion of the right kidney but much
less tumor about it. The right adrenal cannot be located; the
left is free from tumor although there is a necrotic area between
it and the kidney. The liver shows a large nodule of the same
sort of tumor.

SECONDARY TUMORS OF THE KIDNEYS AND ADRENALS

The mouse kidneys seem to be extremely insusceptible to
metastatic invasion by tumors. In our entire series of primary
tumors of mice, now in the neighborhood of 5000 cases, of
which the predominating form is carcinoma of the mammary
gland, we have never met with a metastatic growth from
one of these tumors into the kidneys, with one possible
exception, despite the frequency of pulmonary metastates,
which often replace most of the lung. The only metastatic
carcinomas of the kidney that we have seen have been found

332 MAUD SL YE , H. F. HOLMES AND H. G. WELLS

in four cases of primary carcinoma in the lung (3098, 11777,
12373, 14242). A photograph of the first of these four cases
appears in our paper on lung tumor (52). These carcinomas of
the lung are especially likely to produce extensive metastases,
and the suggestion of some authors that these pulmonary growths
in mice should not be included among the true tumors is evidently
based on ignorance of their character. The fact that they have
furnished our only instances of metastatic carcinoma in the
kidney is sufficient evidence of their true neoplastic character
and the high malignancy of some types.

Although we have seen many instances of widespread metas-
tasis in sarcoma, the kidneys are almost immune from vascular
sarcoma embolism. In our series of 87 cases of sarcoma in mice,
although 23 showed metastasis but one hematogenous secondary
nodule was found in the kidney, this coming from a mediastinal
sarcoma (11791). Since then we have seen one other case of
vascular metastasis of sarcoma into the kidney, from a sarcoma
of the uterus (12058), with a large nodule of the same structure
almost replacing the lower pole of the right kidney, and metas-
tasis in the right ovary. Even the widespread growths of small
round cells, which resemble the condition called lymphosarco-
matosis in man, seem to affect the kidney but little, for among a
considerable number of such cases we have but one with a dis-
tinct metastatic nodule in the kidney (7572), although several
cases show extension from retroperitoneal metastases into the
capsule, and then into the kidney. On the other hand, the
kidneys exhibit extensive diffuse infiltration in leukemia and
massive perivascular growths in pseudoleukemia in mice as in
man.

The malignant retroperitoneal growths, most of which resemble
lymphosarcoma, and the malignant tumors derived from the
adrenals, commonly invade the kidney by direct extension into
the hilum, often very extensively. The sarcomas primary in
one kidney also tend to spread into the hilum of the opposite
kidney. This ready invasion of the hilum of the kidney is a
point of similarity of human and mouse neoplasms.

TUMORS IN KIDNEY AND ADRENAL OF MICE 333

No instance of tumor metastasis into the adrenal has ever
been observed, except possibly in the few cases of widespread
mesotheliomatous growths invading both the ovaries and adrenals,
the origin of which is uncertain.

SUMMARY

In a series of 33,000 autopsies on mice of the Slye stock, dying
natural deaths at all ages, but as far as possible living out their
natural span of life, there have been observed the following cases
of true primary neoplasm arising from renal or adrenal tissues:
First, from the kidney, 16 tumors, classified as follows: 1 carci-
noma, 3 adenomas, 1 hypernephroma, 7 sarcomas, 3 mesothe-
liomas, and 1 sarcoma of the renal pelvis. Second, from the
adrenal, 4 tumors, as follows: 1 cortical adenoma from a mis-
placed inter-renal adrenal rest, 3 mesothelial tumors. Third,
five cases of tumors of the mesothelial structure characteristic
of urogenital anlage neoplasms, but of which the exact origin
could not be determined because of their widespread growth
at the time of death. As these 25 tumors occurred in 33,000
mice presenting not far from 5000 other tumors, they are evi-
dently uncommon tumors of mice, at least in this particular
stock.

It will be noted that in this series there has been no instance
of a mixed renal tumor of the Wilms type, which is so common a
type of renal tumor in man and apparently also in swine. Al-
though inflammatory conditions are very prevalent in the kid-
neys of mice, epithelial tumors are rare, and especially to be
noted is the absence of even a single case of typical malignant
hypernephroma, although one benign growth of this type was
found. No epithelial tumors of the renal pelvis were found,
although there was one case of sarcoma that seemed to take its
origin in the pelvis.

Several instances of malignant retroperitoneal tumors have
been observed, mostly of sarcomatous structure, which usually
invade the kidney. These have not been included in this series,
except two cases in which the structure resembled that of the

THB JOURNAL OF CANCER RESEARCH, VOL. VI, NO. 4

334 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

mesotheliomas, suggesting that the tumor had its origin in mis-
placed rests of the urogenital anlage.

Secondary tumors have never been found in the adrenals,
and but rarely in the kidneys. Although this series includes at
least 3000 cases of mammary carcinoma, often with widespread
metastases in the lungs, we have never seen a secondary carcino-
matous growth in the kidney. The only secondary carcinomas
of the kidney as yet observed are four cases in which the
primary carcinoma was in the lung, thus establishing the true
neoplastic nature of these lung growths. In but two cases have
metastatic sarcomas been seen in the kidney, if we exclude the
numerous cases of invasion of the kidney by direct extension
from para-renal growths.

As to sex: In the entire group of renal and adrenal tumors, we
have equal numbers in males and females, agreeing with the
observation made on other tumors in mice that, in tumors not
peculiar to the sex glands, there is usually little difference in the
incidence in the two sexes.

Differing from the tumors previously studied, coincidence of
other tumors with the renal and adrenal tumors is uncommon.
One mouse in this series had a spindle-cell sarcoma of the thigh.
One mouse had a small, benign papillary adenoma of the lung.
Only two mice had a mammary carcinoma, and one of these
(21663) was a remarkable case, for this animal, when but one
month old, was found with two independent mammary carcino-
mas, and with osteosarcomas in the spinal column and in a rib.
It lived eighteen days more, and at autopsy there was also found
a mesotheliomatous type of growth involving both kidneys.
Except for this unique case there have been practically no in-
stances of malignant tumors in mice less than four months of
age, and few under six months. Most of the renal sarcomas
occurred between the ages of seven months and one year, which is
somewhat earlier than the usual time of appearance of epithelial
growths; this, of course, corresponds to experience with human
neoplasms.

The epithelial renal and adrenal tumors furnished no illustra-
tion of metastasis, but in three cases of sarcomatous or meso-

TUMORS IN KIDNEY AND ADRENAL OF MICE

335

theliomatous growths there was noted involvement of the ad-
jacent lymph-nodes; in two there were pulmonary, in two hepatic,
and in one splenic metastasis, and in one case there were numerous
peritoneal growths. The mesothelial type of growths produced
the most extensive metastasis and the most widespread infiltra-
tion of the body wall.

A review of the literature on renal tumors throughout the
animal kingdom, which constitutes the introductory part of this
article, discloses but six other cases of renal tumors in mice,
all epithelial, and no other adrenal tumors.

REFERENCES

(1) Ttzzer: Jour. Med. Res., 1907, xvii, 166; 1909, xxi, 479.

(2) Haaland: Fourth Sci. Report, Imperial Cancer Research Fund, London,

1911, p. 1.

(3) Murray: Third Sci. Report, Imperial Cancer Research Fund, London,

1908, 73.

(4) Stumpf: Zieglers Beitrage, 1910, xlvii, 571.

(5) McCoy: Jour. Med. Res., 1909, xxi, 285.

(6) Woolley and Wherry: Jour. Med. Res., 1911, n. s. xx, 205.

(7) Bullock and Rohdenburg: Jour. Cancer Res., 1917, ii, 39.

(8) Beatti: La Semana Medica, Buenos Aires, 1917, xxiv, 643.

(9) Loewenstein: Beitr. z. klin. Chir., 1911, lxxvi, 750.

(10) McCoy: Jour. Infect. Dis., 1914, xiv, 53.

(11) Fox: Jour. Path, and Bacteriol., 1912, xvii, 217.

(12) Curtis: Jour. Agric. Res., 1915, v, 397.

(13) Wernicke: Ztschr. f. Krebsforsch., 1911, x, 168.

(14) Burger: Ztschr. f. Krebsforsch., 1914, xiv, 526.

(15) Joest and Ernesti: Ztschr. f. Krebsforsch., 1915, xv, 1.

(16) White: Zool. Soc, Philadelphia, 46th Report, 1918, p. 28.

(17) Fox: 48th Report, 1920; 49th Report, 1921.

(18) Fox: 43rd Report, 1915; Jour. Path, and Bacteriol., 1912, xvii, 227.

(19) Fox: Zool. Soc, Philadelphia, 42nd Report, 1914.

(20) Fox: Jour. Path, and Bacteriol., 1912, xvii, 227.

(21) Fox: Zool. Soc, Philadelphia, 49th Report, 1921.

(22) Seligman: Proc. London Zool. Soc, 1907, p. 173.

(23) Baird: Jour. Cancer Res., 1917, ii, 103

(24) Borell and Masson: Bull, assoc. franchise pour l'etude du cancer, Paris,

1911, iv, 172.

(25) Scott: Jour. Cancer Res., 1917, ii, 367.

(26) Sticker: Arch. f. klin. Chir., 1902, lxv, 616.

(27) McFadyean: Jour. Comp. Path, and Therap., 1899, xii, 139.

(28) Kimura: Gann, 1917, xi, 38.

(29) Trotter: Jour. Compt. Path, and Therap., 1911, xxiv, 1.

336 MAUD SLYE, H. F. HOLMES AND H. G. WELLS

(30) Steinke: Frankfurter Ztschr. f. Path., 1910, v, 167.

(31) Roussy and Wolf: Ann. de m6d., 1921, viii, 464.

(32) Cadiot: Bull, de l'Acad. de meU, Paris, 1921, lxxxv, 396.

(33) Roffo: Rev. d. Inst. Bacterid., Buenos Aires, 1918, i, 333.

(34) Murray: Third Sci. Report, Imperial Cancer Research Fund, 1908, p. 41.

(35) Hamburger: Virchows Arch. f. path. Anat., 1889, cxvii, 422.

(36) White: Zool. Soc, Philadelphia, 46th Report, 1918.

(37) Bland-Sutton: Jour. Anat. and Physiol., 1885, xix, 464.

(38) Wells: Trs. Chicago Path. Soc, 1908, vii, 140.

(39) Pick: Berl. Med. Gesellsch., 1905, ii, 406.

(40) Williams: Natural History of Cancer, p. 11.

(41) Teutschlaender : Ztschr. f. Krebsforsch., 1920, xvii, 322.

(42) Kimura: Gann, 1920, xiv, 10.

(43) Steinke: Frankfurter Ztschr. f. Path., 1910, v, 167.

(44) Smallwood: Anat. Anz., 1905, xxvi, 652.

(45) Fox: Jour. Path, and Bacteriol., 1912, xvii, 217.

(46) Fox: Zool. Soc, Philadelphia, 49th Report, 1921.

(47) Slye, Holmes, and Wells: Jour. Cancer Res., 1920, v, 205.

(48) Ewing: Neoplastic Diseases, 2d ed., Philadelphia 1922, p. 739.

(49) Syle, Holmes, and Wells: Jour. Cancer Res., 1920, v, 205.

(50) Slye, Holmes, and Wells: Jour. Cancer Res., 1919, iv, 207.

(51) de Vecchi and Salmon: Quoted by Ewing: Neoplastic Diseases, 2d ed.,

Philadelphia, 1922, p. 767.

(52) Slye, Holmes, and Wells: Jour. Med. Res., 1914, xxx, 417.

A CRITICAL INVESTIGATION OF THE FREUND-
KAMINER REACTION

LOUIS HERLY

From Columbia University, Institute of Cancer Research, F. C. Wood, Director
Received for publication May 24, 1922
BIBLIOGRAPHY

Human tumors

In 1910, E. Freund and G. Kaminer (1) described certain
phenomena occurring when tumor cells are mixed with the
blood serum from non-cancerous individuals, on the one hand,
or blood serum from cancer patients on the other hand. This
relationship became known as the Freund-Kaminer reaction.
These investigators used tumors obtained at autopsy, because
they found it harder to make cell emulsions with fresh material.
The healthy portions of the tumor were minced, put into a cloth
and pressed out by hand in water containing 0.6 per cent NaCl
plus 1 per cent NaFl; the connective tissue and blood-vessels
remained behind and the cells escaped. The resulting emulsion
was centrifuged and the cells were then washed two or three
times with 0.6 per cent saline solution. The washed cells were
finally suspended in an equal amount of 0.6 per cent saline solu-
tion, and enough NaFl was added to make the solution 1 per
cent NaFl; this proportion of NaFl will preserve the cells for
three weeks. In many cases the cells agglutinated during
washing, so that one washing had to suffice. Organ emulsions
were prepared in the same way, and freed of blood by washing
once with distilled water.

The serum was obtained sometimes fresh, sometimes post-
mortem; sometimes, in later experiments, one-tenth of its vol-
ume of 0.5 per cent NaFl solution was added as a preservative.

337

338

LOUTS HERLY

To perform the test, ten drops of serum and one drop of cell
emulsion were mixed, and a drop of 0.5 per cent NaFl solution
added ; the whole was then well mixed, and the cells in one drop
of this emulsion were counted in a Thoma-Zeiss counting chamber.
The mixture was then incubated in a test tube with a rubber top
at from 37°C. to 40°C, and a second count made. When there
were about 20 cells to a large square, their enumeration was
easy; more than this were difficult to count, and a mixture
with 40 cells to the square, which gave a negative result, gave a
positive result with 20 cells to the square.

Freund and Kaminer found that the serum from noncancerous
subjects destroys most of the cancer cells. The experiment was
controlled by incubating an emulsion of cancer cells in a corre-
sponding amount of 0.6 per cent saline solution containing 0.3 per
cent NaFl; no diminution in the number of cells was found.
Non-cancerous serum did not destroy normal cells, whether
these were from normal persons or from those having cancer.
Heating the serum to 55°C. abolished its power to destroy cancer
cells. The authors concluded that the serum of non-cancerous
individuals must contain a coctolabile substance with carcinoly-
tic powers.

In other emulsions similarly prepared, except that the serum
came from cancer patients, there was no diminution in the cell
content. All tests were made with the cancer cells of one
patient and sera of others.

Freund and Kaminer concluded with the question: "Is the
destruction of carcinoma cells by normal serum due to some
property in the serum, or to some special fragility of the cancer
cell itself?" and they argued that the resistance of carcinoma
cells to carcinoma serum points to some lytic property in normal
serum rather than to a vulnerability of the carcinoma cell.

Freund and Kaminer's paper aroused great interest, because
of the diagnostic possibilities of their reaction, and the experi-
ment was repeated by many other investigators. Thus P. v.
Monakow (2) found lysis absent in 86 per cent of cases of car-
cinoma. A cell destruction up to 10 per cent he considered
negative; between 10 per cent to 25 per cent doubtful; and above

INVESTIGATION OF FREUND-KAMINER REACTION 339

25 per cent positive. He observed that besides cancer cells the
emulsions contain connective tissue cells, which are not de-
stroyed. In his hands only three-fourths of normal sera caused
lysis of cancer cells; one quarter of the normal sera tested acted
like cancer sera and one-fifth of all cancer sera destroyed cancer
cells like normal sera. He, as well as Freund and Kraus, found
that the reaction often failed in sarcoma.

Stammler (3) concluded that the Freund-Kaminer test showed
80 per cent positive reactions.

Ranzi (4), in 1911, in a review of diagnostic reactions for malig-
nant tumors, concluded that the serum of cancer patients con-
tains a substance protective for cancer cells.

Kraus and v. Graff (5) attacked the problem from a different
angle. They knew from the work of Salomon and Saxl that
both in pregnancy and in carcinoma the excretion of oxyprotein
is increased, and that the polypeptids are increased in carcinoma
and in the later months of pregnancy. On the assumption of
some analogy between the two conditions, they tested the serum
of pregnant women and serum from the umbilical cord, and
found that cancer serum as well as that from the umbilical cord
failed to dissolve cancer cells. The serum of pregnancy caused
lysis of cancer cells only inconstantly, while serum drawn after
confinement had no lytic action. They concluded that the
placenta produces alterations in the serum causing a loss of
carcinolytic qualities or the formation of a protective substance.
Rabbit and guinea pig serum acts as normal human serum, and
rat, goat, and sheep serum act as carcinoma serum, the former
being lytic and the latter not.

Kraus, v. Graff, and Ranzi (6) also tested the reaction, and
found that 71.4 per cent of tumor patients examined gave a
positive reaction (lysis below 25 per cent) and 3.5 per cent a
partial lysis. Of patients with other diseases and with benign
tumors, 61.2 per cent gave a negative reaction, 15.3 per cent a
positive reaction, and 23 per cent a partial lysis. Seven cases
that had been operated on and were clinically free from re-
currence gave a negative reaction (lysis over 50 per cent) ; hence
they concluded that the Freund-Kaminer reaction is not based

340

LOUIS HERLY

on an acquired tumor disposition but is due to changes in the
blood serum depending upon metabolic changes in the tumor.

Kraus and Ischiwara (7) tested embryonal cells against cancer
cells, and found that human embryonal cells show even greater
lysis with normal serum than do cancer cells. Furthermore,
embryonal cells were cytolyzed by a cancer serum which had no
action on cancer cells. There was no lysis with fetal blood serum ;
but maternal (retroplacental) serum was lytic.

Rosenthal (8) continued along these general lines, using a
25 per cent emulsion of cells suspended in physiological saline
solution as a control. He believed that only a lysis exceeding
25 per cent can be ascribed to the action of serum. His results
were as follows: Fetal cells plus pregnancy serum gave a lysis
beyond 25 per cent. Fetal cells plus fetal serum showed a
lysis beyond 25 per cent. Fetal cells plus umbilical cord serum
gave no cytolysis.

Arzt and Kerl (9) also have attempted to determine the
diagnostic value of the Freund-Kaminer reaction. They found
in the group of tumors examined 83 per cent correct diagnoses
and 17 per cent wrong. These diagnoses were verified by
biopsy, by operation, or by autopsy. In healthy persons and in
those suffering from non-tumorous diseases the reaction gave
87.5 per cent correct and 12.5 per cent incorrect diagnoses.
Freund and Kaminer, when consulted about these figures,
suggested that the wrong diagnoses might be due to such errors
in technique as: (a) obtaining blood during the height of diges-
tion; (6) admixture of red blood cells with the serum; (c) the use
of insufficiently fresh serum; (d) the use of cells that had lost
their sensitiveness.

Kraus, Ischiwara, and Winternitz (10) reached the decision
that human serum can destroy cancer cells but not normal cells,
and that embryonal cells behave as cancer cells in relation to
adult and umbilical cord serum.

This work attracted the attention of Coca (11) who carried out
more than 150 experiments, some of them in the laboratory
of Freund in Vienna. All his experiments, however, even
those performed in Freund's laboratory, resulted negatively,

INVESTIGATION OF FREUND -KAMINER REACTION 341

no cytolysis being observed in any instance with fresh normal
human, dog, or horse serum. He believed, therefore, that the
cytolytic action of normal sera described by Freund and Kaminer
depended upon some factor as yet uncontrollable, and that it
cannot be made the basis of a differential test for malignant
tumors.

Freund and Kaminer have recently published (12) further
experiments, to show that normal serum and tissues contain
an organic fatty acid able to destroy cancer cells; this they call
"normal acid." Cancer and the serum of cancer patients con-
tain an acid that protects the cancer cell by destroying the
normal acid. Pending an identification of the "normal acid, 5 '
an investigation was undertaken of the saturated dibasic fatty
acids, the series to which they believe "normal acid" belongs,
in order to see which ones are able to destroy cancer cells.

The most recent publication is that of Frankenthal (13), who
found lysis of cancer cells treated with normal serum to beyond
50 per cent; but no lysis of cancer cells treated with the serum of
fourteen patients having non-cancerous diseases; 25 sera of
cancer patients gave 7 (28 per cent) negative results. Her con-
clusions are: (a) normal serum destroys cancer cells, but also
normal liver cells in lesser degree; (b) cancer serum acts differ-
ently from normal serum, in that it protects the cells from de-
struction up to a certain degree; (c) the Freund-Kaminer reaction
is neither constant nor specific and, therefore, of no diagnostic
value for the present.

Animal tumors

Freund and Kaminer used tumors (as stated above) that had
been in the mortuary or the laboratory for indefinite periods of
time, and it is open to speculation what proportion of the cancer
cells so used were alive. They also state that the serum used
in their experiments was sometimes obtained fresh and some-
times postmortem, and one may justly ask whether this serum
might not have been more or less decomposed. Hirschfeld and
Ischiwara, working independently of each other, apparently saw
this objection, for they used freshly drawn animal serum.

342

LOUIS HERLY

Animal tumors also have been employed, their great advantage
being, of course that the viability of their cells can be tested by
inoculation. Thus Hirschfeld (14) made a very fine emulsion
of tumor cells, exposed it to the serum of non-tumor bearing
animals or to that of rats with sarcoma in the proportion of
1 to 6, incubated it for three hours, shaking the mixture every
twenty minutes; and then diluted it with equal parts of 0.6 per
cent saline. One cubic centimeter of this suspension was
inoculated into the right inguinal region of normal animals, and
a series of control animals was inoculated with tumor emulsion
in physiological saline solution. The tumors in the animals
inoculated with tumor cells treated with normal serum were
fewer and smaller than in those inoculated with tumor cells
treated with tumor serum. He concluded that this experiment
proves again that normal serum has a damaging action on the
vitality of tumor cells, while the serum of tumor-bearing animals
lacks this quality.

Finally, Ischiwara (15), working with rat sarcoma demon-
strated that the cell reaction appears relatively late (thirty days)
in the serum of rats, and has some relation to the size of the
tumor.

The reaction has been tested from the refractometric side by
Koritschoner and Morgenstern (16), and by Koritschoner
alone (17).

EXPERIMENTAL

In retesting Freund and Kaminer's experimental conclusions
the present writer has also retested Kraus and v. Graff's observa-
tion (5) that the serum obtained from rabbits and guinea-pigs
acts on cancer cells as does normal serum; i.e., causes lysis,
whereas that from rats, goats, and sheep acts like carcinoma or
placenta serum, i.e., has no appreciable lytic effect. If this
were proved to be correct, then the conclusions drawn from all
the experiments in which normal rat serum was used (Hirschfeld
and Ischiwara) would obviously have to be considered erroneous.

In a series of experiments in which blood was drawn from the
heart of rats and guinea-pigs, and from an ear vein of rabbits,
this observation was tested in the following manner.

INVESTIGATION OF FREUND-KAMINER REACTION 343

The blood, drawn under aseptic precautions, was centrifuged
and the resulting serum was pipetted off, care being taken not to
allow any admixture of red blood cells, since Freund, in a dis-
cussion of reported failures had attributed them to faulty tech-
nique. As another suggested error was the use of serum not
sufficiently fresh, all sera obtained were used immediately.
Freund also attributed failure to the use of blood obtained at
the height of digestion, and to avoid this the experiments now
being described were made with animals which had not been
recently fed.

The tumor used in making the suspension was the Flexner-
Jobling rat carcinoma, since all other experiments along these
lines had been carried out with rat sarcoma in spite of v. Mon-
akow's observation, corroborated by Freund and Kaminer,
that the reaction often fails with sarcoma. The tumors selected
were from thirty to seventy days old, it being generally believed
that transplantability and growth capacity increase, in a general
way, with the age of the tumor. In no case was the skin of the
animal attached to the tumor, nor were any of the growths
ulcerated. Furthermore if an animal appeared sick its tumor
was not used.

The tumors were extirpated under aseptic precautions and the
connective tissue capsule was carefully removed, for previous
observers had stated that the admixture of connective tissue
vitiated the test. Thus the only connective tissue remaining
was the small amount comprising the stroma.

Only the healthy margin of each tumor was used and micro-
scopic sections were made for purposes of checking up each
experiment.

A fine emulsion of the tumor was then made by thoroughly
mincing it with scissors, in order that the serum might come into
contact with all parts of the growth. Various machines devised
for emulsifying tumors were at first used but it was found that the
resultant particles were too coarse and that a much finer emulsion
could be obtained by using very sharp, curved manicure scissors.
All manipulations, of course, were conducted under strictly
aseptic precautions.

344

LOUTS HERLY

This fine tumor emulsion was next added to the serum in the
proportion of 1 to 6, and the suspension was placed in test tubes
and incubated at 37°C, for three hours. Some fresh tumor
emulsion was immediately transplanted, without exposure to
serum, into a series of healthy laboratory rats known to be
susceptible to the Flexner-Jobling carcinoma, in order to prove
the viability of the tumor employed. The amount transplanted
was 0.003 gram fresh tumor emulsion, which was deposited in the
right inguinal region with the usual inoculating needle. The
tubes placed in the incubator were each supplied with a sterile
glass rod and the emulsion was vigorously stirred every half
hour in order to bring all particles of the tumor emulsion thor-
oughly into contact with the serum and so allow the serum to
exert its influence on the largest possible number of tumor cells.
At the end of the three-hour period all supernatant serum was
poured off and the fine particles transplanted into a series of rats.

The protocols of a few experiments selected at random from a
considerable number follow:

Comparison of the number and size of the tumors in figure 1
shows no distinct difference between the control tumors and
those exposed to rabbit serum. In the serum series there is
found one receding tumor; but recession of the Flexner-Jobling
carcinoma is a common event, and does not vitiate the results
of this experiment. One inoculation was negative, but only
about 80 per cent of inoculated animals, under routine conditions,
grow this tumor. One animal died before the first charting.
On the whole, one may say that the animals in both series had
tumors of from 10 mm. to 40 mm. in diameter and that the takes
are entirely comparable in both series.

Reference to figure 2 shows that in both series after forty-five
days there are tumors from 10 mm. to 40 mm. in diameter, the
number of growing tumors being practically 100 per cent, with
the exception of one receding tumor in the serum series.

It is evident from figure 3 that the tumors were more numerous
as well as larger in the serum series than in the controls.

In the next group of experiments the effect of incubation in
guinea-pig serum was studied.

INVESTIGATION OF FRETJND-KAMINER REACTION 345

As may be seen from figure 4, 45 days following the inoculation
there were 100 per cent takes in both series, and there is no
obvious difference between the results in the two groups.

RA38IT SEBUM PLAIN TUMOR EMULSION

DAYS

10 17 24 31 38 45 10 17 24 31 38 45

1 .

13 . Q

14 . •

• ft!

• • f • 15

* ft • • •

i I | If "••••##

• • • • •

_ - • ft •

ti « -

10 CM

FRC

Fig. 1. Experiment nn

89 A

An emulsion of the Flexner-Jobling rat carcinoma, diluted with five parts of
rabbit serum, was incubated for three hours and then transplanted into a series
of 12 rats (nos. 1 to 12). Fresh tumor emulsion was inoculated into a series of
5 rats (nos. 13 to 17) to act as a control.

Reference to figure 5 will prove that there are 80 per cent takes
in the serum group, whereas in the fresh tumor, or control series,
there are only 40 per cent takes, and the tumors are actually
smaller. Clearly the guinea-pig serum exerted no deleterious
effect on the tumor cells.

346

LOUTS HERLY

It can be seen from figure 6 that forty-two days after the
inoculation there were 100 per cent takes in both series and
there is no obvious difference that can be observed in comparing
the results in the two series.

Rabbit and guinea-pig serum may have a different action on
human cells, but for the purpose of the present investigation the
road is cleared by the proof that the serum from neither of these
species has a deleterious effect on the tumor cells employed.

RABBIT SERUM PLAIN TUMOR EMULSION

DAYS

17 24
1 *

31
•

38 45

f0 17 24 31 38 45
1 t % 1 t

• •

•

• •

■ . • • • •

• •

•

• +

• * MM

• •

« • * 3 9 # «

10 CM

w FRO

Fig. 2. Experiment —

89 E

Five rats (nos. 1 to 5) were inoculated with an emulsion of the Flexner-Jobling
rat carcinoma incubated in rabbit serum in the proportion of 1 to 6, and 5 rats
(nos. 6 to 10) with fresh tumor emulsion.

The results of these experiments do not agree with Kraus and
v. Graff's observation that the serum obtained from rabbits and
guinea-pigs exerts a damaging action on cancer cells; for, in
some instances (figs. 3 and 5) the cancer cells subjected to the
action of serum from these animals grew even better than did
cells from the same tumor which had not been subjected to the
action of the serum. We do not attempt to explain the reason
why previous observers have obtained such different results.

The objections to Freund and Kaminer's method have been
already discussed. To obviate these, a series of experiments was

INVESTIGATION OF FREUND-KAMINER REACTION 347

carried out on animal tumors because here both cells and serum
can be obtained perfectly fresh, and the viability of the tumor
cell can be tested by inoculation. Cells from rapidly growing
rat carcinomata (Flexner-Jobling) were suspended in the serum

RABBIT SERUM

DAYS

10 17 24 31

4* 52 59

• • •

Ifl

•

t •

•

• • •

- » •

•

• « #

•

- - . %

•

§ ULCERATED

PLAIN TUMOR EMULSION

DAYS

10 17 24 31

45 52 59

— • • •

• • •

•

• • •

10 CM

F R C

Fig. 3. Experiment — — —
90 K

A series of rats (nos. 1 to 7) was inoculated with an emulsion of the same tumor
incubatedun rabbit serum in the proportion of 1 to 6, and 5 rats (nos. 8 to 12)
with fresh tumor emulsion to serve as a control.

348

LOUIS HERLY

of normal rats, or of rats bearing this tumor, without the inter-
vention of any extraneous substance such as NaCl or NaFl.
It was realized, of course, that only the first half of the experi-
ment is really valid, since the serum of an animal with a trans-

GUINEA PIG SERUM PLAIN TUMOR EMULSION

DAYS DAYS
10 17 24 31 38 45 10 17 2S 31 33 45

1 t •
#

. • • ft • * # • :• • •

,MU * # • • • ♦

• • • • •

. # • ft +

. • • •

• • • • •

• • # ft •

10 CM

FRC

Fig. 4. Experiment — — -
89 A

A series of 10 rats (nos. 1 to 10) was inoculated with an emulsion of the Flexner-
Jobling rat carcinoma incubated in guinea-pig serum in the proportion of 1 to 6,
and a series of 5 rats (nos. 11 to 15) with fresh tumor emulsion for controls.

planted tumor may be entirely different from that of one with a
spontaneous neoplasm. But the extreme rarity of carcinoma
in the rat, and the impossibility of using a sarcoma left no other
course open. Mouse carcinoma was not used because of the
scanty amount of serum obtainable from so small an animal.

INVESTIGATION OF FREUND-KAMINER REACTION 349

8 ® 9

o o

5 •

1 8

2 8 >

SS C 3

9 •

-3 O

03 C

if S

pQ ^

3 _£

03 lO

1 Q

V 03

3 "S

3 *

S O

03 »H

* •s

05 03

-2 as

3 CQ

o c3

§^

C o

3 §■

350

LOUIS HERLY

Nevertheless, it can be shown whether or not normal serum
damages the tumor cell; if it does not, the distinction between
the serum of a normal animal and of one with a spontaneous
tumor vanishes.

GUINEA PIG SERUM

DAYS

7 14 21 28 35 42

1 i | % f

2 - ♦ • • # •

3 . • • • •

4 • • • • •

5 _ . • » % %

6 - • • • • f
' • • • • t •

8 • • 6 « #

9 • M §4

10 . • * • •

PLAIN TUMOR EMULSION

DAYS

14 21

• \

•

35 42

»• m

•

1 i

fit A

14 .

• •

•

* !

•

t *

16 .

1 •

•

• t

•

! 1

• •

20 _

•

FRC

Fig. 6 Experiment ^

A series of rats (nos. 1 to 10) was inoculated with an emulsion of the Flex-
ner-Jobling rat carcinoma incubated in guinea-pig serum in the proportion of
1, to 6, and a series of 10 rats (nos. 11 to 20) with fresh tumor emulsion for
controls.

The experiments were carried out exactly as were those with
rabbit and guinea-pig serum. The blood was obtained from the
hearts of rats bearing Flexner-Jobling tumors from four to eight
weeks old. All animals had large single or multiple tumors,
ranging from 20 mm. to 35 mm. in diameter, none of which were

INVESTIGATION OF FREUND-KAMINER REACTION 351

fixed to the skin or ulcerated. Nor did any of these animals
manifest any evidence of illness.

While the other manipulations necessary to the experiment
were conducted, the serum was placed in the refrigerator at
8°C. Similarly a number of rats without tumors were bled
and the serum also placed in the refrigerator.

A healthy tumor was chosen, its capsule carefully stripped
off, any necrotic material thoroughly scraped out, and the outer
layer of the remaining healthy tumor tissue finely emulsified.
All tumors used in these experiments were from thirty to ninety
days old and microscopic sections of each tumor used were made
and filed away for purposes of identification.

The sera were now taken from the refrigerator and to each was
added this finely minced tumor in the proportion of 1 part of
tumor to 5 parts of serum. A sterile glass stirring rod was added
to each tube, and the tubes were incubated at 40°C. for three
hours. Every half hour each tube was stirred with its glass
rod in order to bring all tumor cells repeatedly into contact with
the serum.

At the end of the incubation period all the serum was poured
off and the fine tumor particles were transplanted into a series
of rats in the manner previously described. A similar dose
(0.003 gram) of fresh tumor emulsion was transplanted in the
same way and at the same site without its having been exposed
to any serum.

The protocols of a few typical experiments follow :

Figure 7 shows that the tumor cells incubated in normal serum
grew better and more uniformly than did those incubated in
tumor serum, and even better than the untreated tumor emulsion
which had not been in contact with any serum whatsoever.

Figure 8 shows again that there was better growth in the
series of animals inoculated with cancer cells incubated in normal
serum, than in the tumor serum series. The results in the un-
treated tumor emulsion series were similar to those obtained in
the normal serum series.

After thirty-eight days, as figure 9 shows, the results in all
three series were practically alike, the tumors varying in diameter
between 7 mm. and 20 mm. and growth being about the same.

TUMOR SERUM KORMAL SERUM

MVS

31 38

1 !

•

13 5

2 .

14 !

•

1 1

3 .

15 .

• 9)

4 i

t
I

•

18 i

1 f

S .

•

e
t

17 •

• •

8 .

•

18 *

•

f €

7 ,

19 *

•

8 .

•

20 f

W T

9 i

21 ■

•

W 0

to ,

•

22 -

•

• t

11 ,

•

23 i

•

• t

12 .

•

24 -

• *

PLAIN TUMOR EMULSION
DAYS

10 17 24 31 38
25 - , 1 g

26 -

" • #

28 - - . t f

10 CM

FRC

Fig. 7 Experiment

83 B

An emulsion diluted in the proportion of 1 to 6 with tumor serum, and in the
same proportion with normal serum, was incubated and then transplanted into
two series of animals, twelve each in number. Into a series of five animals a fresh
undiluted tumor emulsion was transplanted.

352

INVESTIGATION OF FREUND-KAMINER REACTION 353

TUH3.1 SEfflM

DAYS

id t7

31 38

• 1

•

: 1

•

i |

< f

4 i

•

a #

•

• *

- -

•

• #

•

% +

9 +

PUIR TUMOR EMULSION

2S i

28 .

27 .

•

tOtM

„ FRC
Fig. 8 Experiment — — —
84 N

An emulsion diluted in proportion of 1 to 6 respectively with tumor serum and
normal serum was incubated for three hours and then inoculated into two series
of twelve animals each and the fresh tumor emulsion into a series of five animals.

354

LOUTS HERLY

TUMOR SERUM NORMAL SERUM

DAYS

10 17

31 38

4
1

it —
M

•
•

• \

•

ft
•

•

15 '

•

* •

•

• *

•

18 -

* •

•

19 - •

* t

20 - -

•

21 - _

—

•

—

24 _ _

- -

PLAIN TUMOR EMULSION

•

10 CM

F R C

Fig. 9. Experiment —

85 D

An emulsion diluted in the proportion of 1 to C respectively with tumor serum
and normal serum was incubated for three hours and then inoculated into two
series of twelve animals each and a fresh tumor emulsion into a series of 5 animals.

INVESTIGATION OF FREUND-KAMINER REACTION 355

Obviously the serum of normal rats exerts no deleterious
influence upon the cells of the Flexner-Jobling carcinoma, for in
three separate experiments cells incubated in normal serum grew
quite as well as those incubated in serum from rats with trans-
planted tumors. Nor was there any difference in growth between
cells exposed to either serum and those transplanted without
previous manipulation.

CONCLUSIONS

The conclusions to be drawn from the whole series of experi-
ments are equally clear. After having made sure that there was
no difference in the action of either rabbit or guinea-pig serum
on the one hand, and rat serum on the other; and after having
eliminated the sources of error enumerated by Freund and
Kaminer, namely the admixture of connective tissue capsule and
its blood-vessels, the presence of red blood cells in the serum, the
use of old serum or of serum obtained at the height of digestion;
and having, furthermore, used carcinoma (the most sensitive
tumor to this test, according to the originators) instead of sar-
coma, as had been done by previous investigators, we are still
unable to detect any marked difference in the results of inocula-
tion whether tumor serum or normal serum is used. The growth
capacity seems neither increased in the one nor diminished in the
other. We feel justified in stating, therefore, that the value of
the Freund-Kaminer reaction remains at present unproved.
With this proviso, however, that these experiments prove only
that the serum of normal rats is devoid of any deleterious effect
on the Flexner-Jobling rat carcinoma. What might be the case
with serum from a rat bearing a spontaneous carcinoma, we do
not yet know. However, the first half of the experiment seems
sufficient; for since this shows clearly that normal serum has no
harmful effect upon the cancer cell, there can be no difference,
with respect to cytolytic power, between normal serum and that
of an animal with a tumor, be this transplanted or spontaneous.

REFERENCES

(1) Freund and Kaminer: Biochem. Ztschr., 1910, xxvi, 312.

(2) v. Monakow: Miinchen. med. Wchnschr., 1911, lviii, 2207.

356

LOUTS HERLY

(3) Stammler: Ref. Miinchen. med. Wchnschr., 1911, lviii, 1043.

(4) Ranzi: Handbuch d. Tech. & Meth. d. Imm., Ergaenzungsbd., 1911, p. 611.

(5) Kraus and v. Graff: Wien. klin. Wchnschr., 1911, xxiv, 191.

(6) Kraus, v. Graff, and Ranzi: Wien. klin. Wchnschr., 1911, xxiv, 1003.

(7) Kraus and Ischiwara: Wien. klin. Wchnschr., 1912, xxv, 583.

(8) Rosenthal: Ztschr. f. Immunitatsforsch. u. exper. Therap., 1912, Orig.,

xiv, 174.

(9) Arzt and Kerl: Wien. klin. Wchnschr., 1912, xxv, 1821.

(10) Kraus, Ischiwara, and Winternitz: Deutsch. med. Wchnschr., 1912,

xxxviii, 303.

(11) Coca: J. Cancer Res., 1917, ii, 61.

(12) Freund and Kaminer: Wien. klin. Wchnschr., 1919, xxxii, 1105.

(13) Frankenthal: Ztschr. f. Krebsforsch., 1920, xvii, 250.

(14) Hirschfeld: Ztschr. f. Krebsforsch., 1912, xi, 388.

(15) Ischiwara: WTien. klin. Wchnschr., 1913, xxvi, 370.

(16) Koritschoner and Morgenstern: Biochem. Ztschr., 1920, civ, 259.

(17) Koritschoner: Biochem. Ztschr., 1922, cxxix, 605.

IS CANCER A BIOLOGICAL PHENOMENON ?x

SOME HERETIC THOUGHTS ON CANCER
F. G. GADE

Kristiania, Secretary to the Norwegian Committee for Cancer Research
Received for publication June 2, 1922

The primary causes of cancer are still unknown. It seems,
however, more and more evident that the development of cancer
is particularly favoured by certain conditions, viz. chronic local
irritants and old age.

The chronic local irritants are of the most variegated descrip-
tions, belonging to all manifestations of energy — mechanic,
thermic, chemical, radiological; there are also vital (animal)
irritants (Bilharzia, Spiroptera neoplastica s. Gongylonema
neoplasticum, Fibiger), and pathological irritants (precancerous
conditions).

But the very difference in the nature of all these irritants indi-
cates that it is scarcely the irritants themselves but more likely
the reaction of the organism against these, which offers the condi-
tions favourable for the development of cancer. Borrmann and
Ribbert emphasize that, previous to the cancerous proliferation
of the epithelium there is constantly found an inflammatory
alteration of the underlying connective tissue, representing an
alteration of the conditions of life of the epithelial cells.

Very frequently, though, cancer develops in localities where
no irritants or irritation have been noticed.

The influence of age is more obvious; statistics from all
countries here agree. In Norway, the mortality from malignant
tumors, for the years 1901-1915 is shown in table 1.

Of these 32,274 deaths from malignant tumors, only 206, or
0.64 per cent, have occurred in persons under thirty years of age;

1 Paper read in part before the Medical Society of Kristiania, September 21,
1921.

357

358

F. G. GADE

only 1146, or 3.55 per cent, under forty years, though these groups
include 59.27 per cent and 71 per cent respectively of the whole
population.

It is also remarkable that the maximum of deaths from cancer
occurs in the period from sixty to seventy years of age, with nearly
30 per cent (column 6) of all cancer deaths, in spite of this period of
life numbering only 5.96 per cent (column 2) of the population.
And in the period from seventy to eighty years there is still an
increase as to the relative importance of cancer as the cause of

TABLE 1

Cancer mortality in Norway, 1901-1915, distributed in decennial groups

(1)

(2)

(3)

(4)

(5)

(6)

(7)

POPULATION
OF NORWAY 1910

CANCER MORTALITY IN NORWAY

AGE

Number of
individuals

Per
hundred
of t he pop-
ulation

Total
from
1901-1915

Average
yearly

Per hun-
dred of

all cancer
deaths

Yearly

per
100,000
living

years

1-20

1,063, 262

45.10

2.53

0.12

0.024

20-30

334, 091

14.17

168

11.20

0.52

0.34

30-40

275, 800

11.69

940

62.66

2.91

2.27

40-50

225, 181

9.56

3, 226

215.06

9.99

9.55

50-60

194,109

8.24

6, 925

461.66

21.45

23.73

60-70

140, 579

5.96

9, 658

634.85

29.93

45.80

7O-S0

83, 825

3.56

8, 462

564.15

26.22

67.30

80-90

32, 551

1.38

2, 716

181.00

8.41

55.60

3,065
5,327

0.13

141

9.4

0.45

30.67

No age reported

0.19

Total

2, 357, 790

99.99

32,274

2151

100.00

9.15

death. After the age of eighty years there seems to be some
decline in the relative cancer mortality; but this is without doubt
only apparent, numerous deaths from cancer in the latest decen-
niums of life being covered by the diagnosis of senile debility as
well as by other intercurrent causes of death.

Former investigations by the Norwegian Committee for Cancer
Research as well as the recent ones by Dr. J0rgen Berner make it
very probable that the frequency of can cor continually increases
with increasing age.

IS CANCER A BIOLOGICAL PHENOMENON

359

The table shows that the ages above thirty to forty years are
affected by cancer proportionally 226 resp. 66 times as often as
the ages below the same years.

Age must consequently be of the greatest importance for the
development of cancer, and Bashford's saying, that " cancer is a
function of age" contains a great amount of truth.

It must therefore be one of the nearest and most important
objects for the cancer research work to investigate this influence
of age.

Surely it has been hinted that age just gives the chronic irritant
sufficient time to act. But this explanation of the influence of
age is hardly satisfying. We do not know the time necessary
for any irritant to produce cancer. But suppose that even ten
or twenty years are required, there will still be every opportunity
for an irritant to unfold its activity in many more thousands of
cases in the twenty to forty years group, than in the older groups,
but cancer is very rare at the younger age. And some of
the irritants, for example, the z-ray, seem to be able to pro-
duce cancer after a proportionally short time and in younger
individuals.

It seems more natural to attribute the influence of age upon the
development of cancer to the change of metabolism in the aged.
This change alters conditions of life for cells and tissues, and a
local irritant may add directly and indirectly to these alterations
and accentuate the influence of senile metabolism.

The influence of age on the development of cancer is so pro-
nounced that it seems to signify a biological law. That cancer
occasionally may develop in younger individuals makes no differ-
ence as to the regular influence of age. It is not the mere number
of years of life that constitutes "age", but the individual's physi-
ological condition. This senile condition may be postponed in
many aged individuals; on the other side a senilitas praecox can
be found in younger individuals, as we see it in the arteriosclerosis
with young neurasthenics.

Unfortunately we do not know enough about the metabolism
of age to be able to form any precise conception of its influence
upon cancer formation. "Geriatry" has not been studied as
thoroughly as pediatry.

360

F. G. GADE

Amongst the more important somatical peculiarities of age may-
be mentioned: diminution of the total mass of blood, reduction
of produced and needed calories, lower temperature of the body,
greater amount of cholesterin in the fat cells, increased total
nitrogen in the blood serum, reduction of the number of cells
in the connective tissue and of their regenerative power, atrophy
of the elastic tissue, of lymph nodes, spleen, Peyer's lymphatic
noduli, and of the bone marrow. Then, also, a change and
diminution of the functions of the endocrine organs.

The senile changes appear at different times in different indi-
viduals as well as in the different organs of the same individual.
The senile changes can occupy the entire organism (general
senility) or only certain organs or parts of these (local or partial
senility). Local senility can be found in the female generative
organs (ovaries, uterus, and mammae), where physiological
activity and involution occur years before the manifestation
of the individual's general senility.

In accordance with this fact we see that cancer of the female
reproductive organs appears earlier in life than cancer in the
other organs, the great majority of deaths here occurring between
forty and sixty years of age (Norwegian statistics), in contrast
to deaths from cancer in other organs, where the greatest numbers
are to be found from sixty to eighty years of age, twenty years
later. The chorionepithelioma malignum has also its greatest
frequency in individuals above forty years of age (Anton Sunde).

At present we may assume that local chronic irritants, and old
age, produce the altered conditions of cellular life which entail
cancer.

The real nature of these conditions is still unknown to us. But
it may be worth remembering, that Freund and Kaminer ten
years ago have found in the blood serum and tissue fluids of healthy
(non-cancerous) persons a well defined fatty acid ("normal acid"),
soluble in ether and having the power of dissolving cancer cells.
This acid, that has some resemblance to the decamethylen-
carbonic acid (COOH)2 (C2H4)5, of the succinic acid series, is not
to be found in blood or tissue serum from cancerous individuals,
whose blood has a higher alkalinity than that of non-cancerous
individuals (Maud Menten).

IS CANCER A BIOLOGICAL PHENOMENON 361

We know from the experiments of Jacques Loeb and others that
the division and multiplication of cells can be influenced by very
small changes of their surrounding nutritive mediums. Similar
facts are known from bacteriology.

Robertson and Burnett have been able to accelerate the growth
of cancerous tumors, inoculated upon rats, by injecting the animals
subcutaneously with cholesterin (and with tethelin), substances
which in themselves do not produce cancer in the said animals.
Luden and Bloor have also stated that the blood of cancerous
patients contains more cholesterin than normally. That the fat
cells in old individuals are richer in cholesterin than those in
younger persons has been mentioned above.

Ferd. Blumenthal noticed that the growth of inoculated cancer
in rats was accelerated when the animals got a surplus of potas-
sium in their food, but retarded by a surplus of calcium.

Direct and local influences upon the growth of the epithelium
are known since the experiments of Fischer and others with
Scharlach R (and other similar substances), the application of
which on the skin of rabbits produced a lively proliferation of the
epithelium. And Yamagiwa and Ischikawa, and now lately
Fibiger have through similar application on rats succeeded in
producing growth possessing the essential qualities of the genuine
carcinoma.

The local irritations and the old-age changes of metabolism
entail both altered conditions of life for the cells, consequently
also a change of their general manifestation of life.

But such changes in general manifestations of life are just what
we find as criteria for the transformation of a normal body cell
into a cancer cell: changes in outer shape and structure building
power, changes in demand for nourishment, in speed of growth
and proliferation, in chemical structure, enzymes and secretions
(cachectic power), in sensibility towards electrical influences
(x-rays), and independence of soil (metastases).

These changes of essential biological qualities in the cancer cell
are just what v. Hansemann has designated under the name of
anaplasia (transformation), a phenomenon hitherto considered as
unique, unparalleled.

362

F. G. GADE

But is such a transformation really so unique in the living
nature?

About twenty years ago de Vries introduced into biology the
phenomenon of mutation, the comparatively quick changes of the
characteristics of species in living beings, in contrast to the very
slow changes admitted by the Darwinistic principles of evolution.
This mutation, first noticed by de Vries in a plant (Oenothera), has
later been described for unicellular organisms, bacteria (Weisser,
Massini), algae, spirochaetae, trepanosomas, infusoria, and
amebas. Thj0tta and Eide have described a paratyphoid strain
that, under culture for about a month, essentially changed its
character morphologically and physiologically, loosing its flagelli
and motility and changing into a capsulated mucus-forming
bacillus; which shape it preserved through all the consecutive
generations.

Gurney Dickson, in his monograph on the Transmutation of
Bacteria, maintains that mutation is an adaption to altered
conditions of life; it can be more or less pronounced and may
include very different qualities of the bacteria: shape and size,
affinity to dyes (chemical and physical properties of the proto-
plasm) , ability of growing on a new soil, resistance against external
influences, virulence, and production of ferments. Most fre-
quently mutation occurs in cultures of old date.

There seems thus to be a considerable parallelism between muta-
tion and the phenomena accompanying the transformations of a
normal body-cell into a cancer cell. And the mutated cells, like the
cancer cell, will transfer the newly acquired qualities to their
future generations.

Through all these changes in its fundamental biological
qualities the cancer cell has turned out to quite a new cellular
being, foreign or even hostile, to the organism in which it has
developed. Authors like Bashford, Abderhalden, Ribbert, Butlin
and others call the cancer cell a parasite. Butlin even described
it as a new and special species of beings under the name of
unicellula cancri.

The idea that the transformation of the normal cell to a cancer
cell may be a mutation, can certainly not explain the nature of

IS CANCER A BIOLOGICAL PHENOMENON

363

cancer, the nature of mutation being still enigmatical, but it
brings the cancer question in a line with other biological phenom-
ena, which may be examined and studied on a broader base; and
results from other spheres of biological research may be applied
upon the cancer questions.

Considering the dominating influence of age on the develop-
ment of cancer, it seems necessary for the future cancer research
to direct the closest attention to the metabolism of the aged.
Here will be one of the future battlefields against cancer. The
British x-ray expert, Morley Robert, thinks it essential here to
produce a mental and bodily youth in the individuals. The war-
cry of the medical profession in its struggle against cancer will be :
Keep humanity young !

THE RELATION OF MUSCULAR ACTIVITY TO
CARCINOMA1

A PRELIMINARY REPORT

IVAR SIVERTSEN and A. W. DAHLSTROM

Minneapolis, Minnesota

Received for publication May 20, 1922

Carcinoma in men, especially retired farmers, has been frequently
seen by us in the past few years. These patients, men with large,
well developed bodies of the " frontiersman" type, usually gave a
history of having retired from the farm in good health, a few years
previously, to enjoy a well deserved rest. On the other hand
we have often talked with farmers, who for one reason or another
still shouldered the responsibilities of the farm, who were well
and working hard though sixty, seventy, or even eighty years of
age. This observation brought us to consider muscular activity
in relation to carcinoma.

There have been no deaths from carcinoma among the Turn
Teachers (1) in the past thirty years. Thirty deaths from all
causes occurred in this period, and we should expect at least three
deaths from carcinoma. The Turn Teachers are of necessity
active in a muscular sense. While the list is not large, the excep-
tion to the rule is interesting.

We have repeatedly noted that patients, clinically diagnosed as
precancerous, improved in a most satisfactory manner on a pre-
scription for increased daily exercise, as rope skipping for women,
golf and brisk daily walks for men.

Muscular activity stimulates metabolism and retards the
approach of senescence and decay. The tuberculosis problem
has been greatly simplified by the recognition of the value of rest,
sunshine, wholesome food and fresh air. We believe that the

1 Read before the Hennepin Count}- Medical Society March 6, 1922.

365

THE JOURNAL OF CANCER RESEARCH, VOL. VI, NO !

366

IVAR SIVERTSEN AND A. W. DAHLSTROM

carcinoma problem may also be simplified by the recognition
of the value of muscular activity in those periods of life most
affected by carcinoma, those periods of life when we are prone to
"let down" and check our pace. If muscular activity has an
intimate relationship to carcinoma, the value of it will be most
apparent in prophylaxsis, the field which to date promises the
most in the attack on this malady.

In a perusal of the literature we are unable to find any direct
reference to the relation of muscular activity to carcinoma.2
Hoffman (2) shows that carcinoma is world wide, is a disease of
maturity, affecting both sexes and seriously affecting all civilized
nations and races. Observations are quoted and statistics are
presented showing that carcinoma is more frequent among the
unoccupied, among people of independent means, among the
leisured and professional classes, among the overnourished, well
housed, well fed, and among those who live in the warmer climates.
Carcinoma is less frequent among the people who are actively
engaged in a gainful occupation, among the hard working classes,
among those who live in large industrial centers, among peoples
predominatingly poor and of necessity frugal, living on an alimen-
tation just sufficient for their actual needs, and among those who
live in the colder climates where the struggle for existence is more
difficult. Reading between the lines one common factor is noted,
that of the varying degree of muscular activity. The vast
majority of people follow the line of least resistance and those of
independent means, the leisured and professional classes, and
those who live in the tropics are not compelled to do manual labor
as a means of gaining a living and hence are less liable to be active
in a muscular sense. Attention must be called to Dublin's (3)
conclusion with reference to the experience of the Metropolitan
Life Insurance Company for 1914, 1915, 1916, that "the cancer

1 It is interesting to us to find that in another part of the world similar ideas
have been advanced almost simultaneously. Cherry, of Melbourne, on March 1,
1022, read a paper on A Theory of Cancer (published in Med. Jour, of Australia,
April 22, abstracted in J. A. M. A., 1922, lxxix, 245) in which he attributes
the recent increase of cancer to deficient muscular activity, overnutrition,
and lessened skin activity, as a result of modern living conditions of civilized
peoples.

RELATION OF MUSCULAR ACTIVITY TO CARCINOMA 367

mortality rate at ages where the cancer rate is significant,
decreases as we go up the economic scale." To reconcile the
statistics presented by Hoffman with those of Dublin one must
admit that other factors than the economic status are present.

Ewing (4) states that " statistics show that sanitary measures
which control infectious diseases exercise no such power over
carcinoma, which is somewhat conspicuous by a relative failure to
attack the poor, the overworked, the underfed, the savage, but
chooses a notable proportion of its victims among the well-to-do,
the well nourished, the well protected against infectious diseases,
and the indolent. "

Ewing has noted also that carcinoma seems to be as widespread
as are animal species. The mortality from malignant tumors
among pet dogs is reported to be surprisingly high. Forms of
carcinoma in horses, cattle, and swine are relatively common,
carcinoma of nearly every organ of fowls has been reported, mice
and rats are subject to carcinoma and made use of extensively
in the laboratory. It is very interesting to note that he too calls
attention to the rather sharp fall in carcinoma incidence when
one takes up the consideration of carcinoma in wild animals.
In this respect one must also consider that wild animals have not
been subjected to uniform close study, autopsies have been less
frequent, and violent deaths have been more common than among
the domesticated animals. McCoy (5) reports autopsy findings
of 100,000 rats (95 per cent of which were Mus norvegicus) caught
during the San Francisco rat campaign about fifteen years ago.
Of these rats 103 were found to have tumors, of which 10 proved
to be carcinoma. Thus he found that about one rat for every
thousand had a new growth, and only 1 in 10,000 had a carcinoma,
the remainder being benign tumors or sarcomata. In addition
to the factors incident to the study of carcinoma among wild
animals, the necessary muscular activity of the animal in its
natural wild state may have a bearing on the relative infrequency
of reported carcinoma.

Carcinoma has been observed under certain conditions among
artificially bred fish (6). It has been clearly shown that fish
kept under artificial conditions are more susceptible to the

368

IVAR SIVERTSEN AND A. W. DAHLSTROM

development of carcinoma than are those living under natural
conditions. We believe that the enforced inactivity of the fish
in ponds and hatcheries is a potent factor in the development of
carcinoma. The dietery factor must also be considered.

The present day labor saving machinery, convenient means of
transportation, and ultra-conveniences of modern life all tend to
make us less active physically, and it appears to us that the
increasing curve of carcinoma closely approximates the inaugura-
tion of the Age of Machinery.

The recent researches into the finer biochemical changes
attending the metabolism of carcinomatous patients point to the
development of carcinoma as an individual biological reaction to
chronic irritation, governed possibly by the previous state of
metabolism. It would seem that we must look to the biochemist
for information as to the exact cause of the reversion of the adult
cell to fetal characteristics. The researches of Freund and
Kaminer (7), Rosenthal (8), Neuberg (9), Koritschoner and
Morgenstern (10), and others are very interesting in this
connection.

This opportunity is taken for the presentation of a working
hypothesis: That human carcinoma ma/ be the reaction to and
the result of chronic irritation of adult epithelial tissues bathed in
body fluids altered by certain metabolic products as a result
of deficient muscular activity.

In advancing the above hypothesis constructive and destructive
criticism is invited, as by intensive cooperative effort sufficient
data pro and con will be obtained definitely to prove or disprove
our contention.

An investigation in an experimental way of some of the phases
of muscular activity and its relation to carcinoma is being con-
ducted which will be reported in the near future. We refer
especially to the investigation of the effect of muscular activity
on experimental carcinoma in mice. We are also preparing to
ascertain the effect on human carcinoma of horse serum obtained
after graduated amounts of muscular activity.

In an effort at this time to gain some concrete evidence bearing
on the subject, a study of carcinoma in Minnesota for the past

RELATION OF MUSCULAR ACTIVITY TO CARCINOMA 369

three years was undertaken. Access to original certificates of
death was obtained through the courtesy of the Bureau of Vital
Statistics, Minnesota State Board of Health, there being no
compilations available. Among 86,838 records examined 6,351
deaths were ascribed to malignant disease for the years 1918,
1919, and 1920. Sarcoma was named as cause of death in 452
records. There were 3135 deaths from carcinoma among males,
and these were used as the basis for the present study. Car-
cinoma among females was not studied at this time because of the
difficulty in estimating the amount of muscular activity rela-
tive to occupations of this sex. The deaths were tabulated
according to occupation and the occupations divided by us into
six groups according to our idea of the muscular activity
necessary to that occupation. It is to be regretted that no
scientific criterion was available for this classification. Some
method as suggested by Waller and DeDecker (11) might be
applicable if standardized by sufficient usage. We have tried to
use our best judgment in each case. A special report was
obtained from the Bureau of Census on the number of persons
engaged in each occupation based on the Census of 1920. The
deaths in each group for three years were totalled, averaged, the
number of persons engaged calculated and the death rate com-
puted on the basis of 100,000 persons engaged to bring each group
to a common ground for comparison. The death rates are in a
measure not absolutely reliable because of the short period of
time under consideration and the present day laxness of the
average undertaker in completely filling out certificates of death.
We grouped the occupations as follows: Group 1. Occupations
involving great muscular activity. In this group are included
stonecutters, blacksmiths, boilermakers, moulders and the like.
Group 2. Occupations involving moderate amounts of muscular
activity, as carpenters, cabinet makers, common laborers, masons,
plumbers, and the like. Group 3. Includes occupations involv-
ing medium amounts of muscular activity as foremen, millers,
brewers, agents, printers, etc. Group 4. Occupations involving
small amounts of muscular activity, as the professions, office
workers, and the like. Group 5. The farmers, because of the

370

IVAK SIVERTSEN AND A. W. DAHLSTROM

seasonal character of their work in Minnesota, were classed alone.
Group 6. This group includes those not actively engaged in any
gainful occupation, students, idiots, epileptics, inmates of insane
hospitals, and those who have retired from gainful occupations.
Hoffman (12) reports the following statistics:

Death rate per 100,000 males, England-Wales, thirty-five years of age and over

The differences in mortality are very striking and we believe
they may be explained on the basis of muscular activity.

In Minnesota for the years 1918, 1919, 1920 we have found the
death rate per 100,000 males as follows:

1. Those 21 years of age and over:

Active (groups 1, 2, 3, 4, 5) 103.5

b. Inactive (group 6) 510.0

2. Those 45 years of age and over:

a. Active (groups 1, 2, 3, 4, 5) 285.8

b. Inactive (group 6) 681.0

The figures for Minnesota are seen to parallel roughly those for
England and Wales (see chart 1).

In a consideration of the age class 21 years of age and over we
have found the death rate from carcinoma lowest in group 1
(in which the necessary muscular activity is greatest) and highest
in group 6 (in which the necessary muscular activity is least).
There is a stairstep gradation between the extremes. The ratios
between the death rates in the various groups may be thus
expressed: I:II:III:IV: V: VI: 1 : : 1: 1.7: 2.1 : 2.3: 3.6: 11.8 (see
chart 2, table 1).

To ascertain the ratios in the various age classes we have
subdivided the active groups:

A. Those twenty-one to forty-four years of age. In this age
class the ratios may be expressed thus: I: II: III: IV: V: : 1 : 1 .4:
1.5:2.2:2.0 (see table 2, chart 3).

1890-1892

1900-1902

All males

Occupied

Not occupied

165.5
146.0
359.0

210.3
180.0
503.8

RELATION OF MUSCULAR ACTIVITY TO CARCINOMA 371

DEATH RATES PER 100,000

MALES

Engl and -Wales

Minnesota

1890-1892 1900-1902

1918-1919-1920

Chart 1. Columns lettered A denote occupied; B, not occupied; columns not
numbered denote those thirty-five years of age and over; numbered 1, twenty-
one years of age and over; 2, forty-five years of age and over.

372

IVAR SIVERTSEN AND A. W. DAHLSTROM

Chart 2. Columns Numbered I, II, etc., Refer to Occupation Groups 1,2, etc.

table 1

Death rates per 100,000 males, twenty-one years of age and over, Minnesota, 1918,

1919, 1920

GROUP

NUMBER OF PERSON'S

DEATHS

DEATH RATE

I 1 I

4r>,73'3

20.0

43.0

168,460

128.0

75.8

o
O

84,523

77.0

91.1

211,824

213.0

100.8

170,483

267.7

156.6

1,2,3,4,5

682,026

706.4

103.5

01,503

330.0

510.0

RELATION OF MUSCULAR ACTIVITY TO CARCINOMA 373

B. Those forty-five to sixty-four years of age. The ratios
existing are I:II:III:IV:V: : 1:1.8:2.1:2.0:2.0 (see chart 4,
table 3).

C. In those sixty-five years of age and over we have found the
difference less striking as might be expected for this age class,
the ratios being: I:II:III: IV: V: : 1: 1.5:1.1:1.4:1.9 (see table
4; chart 5).

D. Considering those of another age class, from twenty-one
to sixty-four years of age, we find there is progressive increase in
death rate with progressive diminution in amount of necessary
muscular activity. The ratios may be expressed: I: II: III: IV:
V : : 1 : 1 .6 : 2.3 : 2.4 : 3. 1 (see table 5, chart 6) .

In all the age classes considered the death rate from carcinoma
among those actively engaged in a gainful occupation is roughly
inversely proportional to the amount of muscular activity
necessary to that occupation. The only out and out exception to
the rule is in group 5, the farmers, for which we lack a satisfactory
explanation except that the seasonal character of the occupation
in Minnesota may have a bearing, in that during the spring and
fall seasons the work is very heavy, while in the winter and sum-
mer, especially in winter, the work is very light.

The experience of the Industrial Department of the Metro-
politan Life Insurance Company (13) discloses that cancer causes
the following percentage of total deaths for the following occu-
pations :

Group 1. Blacksmiths 7.6 per cent; miners 4.6; moulders 3.9;
average for the group 5.9 per cent.

Group 2. Laborers 5.5; longshoremen 4.3; machinists 4.5;
masons 6.5; average 5.2 per cent.

Group 3. Teamsters 3.6; printers 2.7; painters 4.1; track
laborers 4.7; average 3.8 per cent.

Group 4. Clerks 3.1; cigarmakers 5.2; average 4.2 per cent.

Group 5. Farmers 7.6 per cent.

The above figures do not indicate any direct influence of the
occupational disease factor.

DEATH RATKS PER 100,000 M*LES
21 to 44 YEARS OP AGE,
MINNESOTA 1918-1919-1920.

DEATH R.'TES PER 100,000 MALES
45 to 64 YEARS OP AGE,
MINNESOTA 1916-1919-1920 .

Chart 3

Chart 4

DEATH RATES PER 100,000 MALES
65 YEARS OF AGE AND OVER,
MINNESOTA 1918-1919-1920.

Chart 5

Chart 6

Chart 3. Columns Numbered I, II, etc., Refer to Groups 1, 2, etc.
Chart 4. Columns Numbered 1, II, etc., Rkfer to Groups 1, 2, etc.
Chart 5. Columns Numbered I, II, etc., Refer to Groups 1, 2, etc.
Chart 6. Columns Numbered I, II, etc., Refer to Groups 1, 2, etc.

374

RELATION OF MUSCULAR ACTIVITY TO CARCINOMA 375

TABLE 2

Death rates per 100,000 males engaged, twenty-one to forty-four years of age, Minne-
sota 1918, 1919, 1920

GROUP

NUMBER OF PERSONS

DEATHS

DEATH RATE

34,587

2.7

7.8

128,233

14.3

11.1

58,362

6.7

11.5

141,902

24.7

17.4

93,661

15.7

16.7

TABLE 3

Death rates per 100,000 males engaged, forty-jive to sixty-four years of age, Minne-
sota, 1918, 1919, 1920

GROUPS

NUMBER OF PERSONS

DEATHS

DEATH RATE

11,057

11.3

102.2

4,373

65.3

189.9

22,955

51.0

221.1

61,321

126.0

205.5

66,168

136.0

205.5

TABLE 4

Death rates per 100,000 males engaged, sixty-jive years of age and over, Minnesota,

1918, 1919, 1920

GROUPS

NUMBER OF PERSONS

DEATHS

DEATH RATE

1,092

6.0

549.4

5,854

48.3

825.0

3,206

19.3

602.0

8,601

65.0

755.7

10,654

115.7

1085.0

TABLE 5

Death rates per 100,000 males engaged, twenty-one to sixty-four years of age, Minne-
sota, 1918, 1919, 1920

GROUPS

NUMBER OF PERSONS

DEATHS

DEATH RATE

45,644

14.0

30.6

162,606

79.7

49.0

81,317

57.7

70.9

203,223

148.7

73.1

159,829

151.7

94.9

376 IVAR SIVERTSEN AND A. W. DAHLSTROM

A tabulation of the ages at death in the different occupation
groups discloses a rather close approximation of extremes and
averages so it appears the differences in death rates shown above
can not be wholly explained on a basis of age incidence; that is,
there is not sufficient difference between the ages of those in the
different occupation groups to explain the ratios found to exist
between the death rates in the various groups of occupations
(see table 6).

TABLE 6

Ages at death, males, Minnesota, 1918, 1919, 1920

GROUPS

NUMBER OF DEATHS

EXTREMES

AVERAGE AGE AT DEATH

24-81

57.3

379

29-91

60.3

201

29-83

57.7

667

24-88

56.8

878

25-90

62.2

984

27-103

70.4

SUMMARY

In a preliminary report attention is called to a new factor in
carcinoma etiology. Carcinoma has been frequently observed in
retired farmers, seldom in active individuals. Precancerous
patients have improved clinically with increased muscular
activity. Attention is called to observations of less carcinoma
among those of necessity physically active. The reported
incidence among domesticated animals is greater than in wild
animals. A working hypothesis is advanced: That human car-
cinoma may be the reaction to and the result of chronic irritation
of adult epithelial tissue bathed in body fluids altered by certain
metabolic products as a result of deficient muscular activity.
From a study of carcinoma deaths among males in Minnesota for
three years it appears that the death rate in those who are active
is greatly exceeded by the death rate in those who are in-
active. From a study of the death rates of those who are
actively engaged in a gainful occupation it appears that the death
rate is lowest in those occupations involving the greatest amounts
of necessary muscular activity, and is highest in those occupations
involving the least amounts of muscular activity. The age

RELATION OF MUSCULAR ACTIVITY TO CARCINOMA 377

incidence factor of the cases studied does not explain the varia-
tions shown. The figures from the experience of the Metropoli-
tan Life Insurance Company do not show any marked influence
of occupational disease on carcinoma death rate. Additional
study and accumulation of data are necessary to establish
definitely the status of the relation of muscular activity to
carcinoma.

CONCLUSIONS

1. Carcinoma constitutes a serious menace to the adults of
all civilized races.

2. It appears that the recent increase has accompanied the
advent of the Age of Machinery.

3. The reported incidence among the lower animals appears
inversely proportional to the degree of muscular activity necessary
to the existence of the animal.

4. The death rate in males actively engaged in a gainful
occupation is less than the death rate in those not actively engaged
in any gainful occupation.

5. The death rate among males actively engaged in a gainful
occupation is inversely proportional to the degree of muscular
activity necessary for that occupation.

6. A new factor in carcinoma etiology is proposed.

7. A working hypothesis is suggested.

REFERENCES

(1) Personal Communication from the Secretary of the Turn Teachers Mutual

Benefit Association.

(2) Hoffman, F. L.: The Mortality from Cancer Throughout the World.

Newark, 1915.

(3) Dublin, L. I.: Mortality Statistics of Insured Wage Earners and their

Families. Metropolitan Life Insurance Company, 1919, Chapter ix.

(4) Ewinq, J.: Animal Experimentation and Cancer. Defense of Research,

Pamphlet iv, A. M. A., 1911, pp. 4-6.

(5) McCoy, G. W.: A preliminary report on tumors found in wild rats. Jour.

Med. Res., 1909, xvi, 285.

(6) Bainbridge, W. S.: The Cancer Problem. New York, 1914, p. 55.

(7) Freund, E., and Kaminer, G.: Reactions between tumor cells and blood

serum. Biochem. Ztschr., 1910, xxvi, 312.

(8) Rosenthal, E.: The biological parallel between cancer and fetal cells.

Ztschr. f. Immunitatsforsch. u. exper. Therap., 1912, xiv, 174.

378

IVAR SIVERTSEN AND A. W. DAHLSTROM

(9) Neuberq, C: Chemistry of tumors. Biochem. Ztschr., 1910, xxvi, 344.

(10) Koritschoner, I. R., and Morgenstern, O.: Refractometric studies of the

Freund-Kaminer reaction. Biochem. Ztschr., 1920, civ, 259.

(11) Waller, A. D., and DeDecker, G.: The physiological cost of muscular

work. Brit. Med. Jour., 1921, i, 669; 1921, ii, 627.

(12) Hoffman, F. L.: The mortality from cancer in different occupations.

Ztschr. d. osterr. Sanitatswesen, 1918, supplement to 1-26, pp. 33-42.

(13) Dublin, L. I.: Causes of death by occupation. Bull. U. S. Bureau Labor

Statistics, No. 207, U. S. Government Printing Office, 1917.

INDEX

Activity, muscular, The relation of, to carcinoma. A preliminary report 365

Adenoma, atypical, An, of the pancreas originating in islet tissue 277

Adrenal, Primary spontaneous tumors in the kidney and of mice 305

Age, Teratomas and their relation to 261

Albino rats, The influence of certain diets upon tumor susceptibility and

growth in (summary) 100

Animal tumors, transplanted, The effect of combined heat and radiation

upon (summary) 101

Animals, immune, Effect of blood from, upon transplantable tumors -25

Bagg, Halsey J., Interstitial injections of an active deposit of radium emana-
tion in a rat carcinoma (summary) 104

Benedict, Stanley R., and Sugiura, K., The influence of certain diets upon

tumor susceptibility and growth in albino rats (summary) 100

Bernhard, A., Rohdenburg, G. L., and Krehbiel, O. F., Further investigations
of disturbances of blood sugar equilibrium in their relation to neoplasia. . 223

Biological phenomenon, Is cancer a. Some heretic thoughts on cancer 357

Biology, tumor, A phase of (summary) 96

Bladder, cancer of the, Radium in (summary) 190

Blood, Effect of, from immune animals upon transplantable tumors 25

sugar equilibrium, Further investigations of disturbances of, in their

relation to neoplasia 223

, The protein content of the whole, and plasma in cancer 127

Bovie, William T., Further experiments on sensitization to heat by radiation

(summary) 193

Breast, cancer of the, A few late results after the radical operation

for (summary) 88

Bulkley, L. Duncan, Proofs of the constitutional nature of cancer (summary) 185

Burrows, Montrose T., Problems in cancer research 191

Burrows, Montrose T., Problems in cancer research 131

Cancer and parasite 257

, Is, a biological phenomenon? Some heretic thoughts on cancer 357

, mortality, Is, increasing? 251

, of the bladder, Radium in (summary) 190

, of the breast, A few late results after the radical operation for (summary) 88

, operations, End -result reports of (summary) 187

, Proofs of the constitutional nature of (summary) 185

, research, Problems in 191

, research, Problems in 131

, Research, Proceedings of the American Association for, Fourteenth

Annual Meeting 175

379

380

INDEX

Cancer, Research, Proceedings of the American Association for, Thirteenth

Annual Meeting 87

, The protein content of the whole blood and plasma in 127

Carcinoma, A comparison of the growth of mice which ultimately develop,

with the growth of mice which do not develop carcinoma 7

, Primary spontaneous squamous cell, in mice. Studies on the incidence

and inheritability of spontaneous tumors in mice 57

, rat, Interstitial injections of an active deposit of radium emanation

in a (summary) 104

, squamous-cell, The occurrence of, in mice (summary) 180

, The influence upon the growth of transplanted Flexner-Jobling rat, of

hydrogen ions and of various salts in different concentrations 285

, The relation of muscular activity to. A preliminary report 365

Cell, fat, A study of a lipomyxosarcoma with comments upon the origin of

the 109

Chromatophores, Inoculation of sarcomatous tumors into negro fowls, with

special reference to the significance of 31

Dahlstrom, A. W., and Sivertsen, Ivar, The relation of muscular activity to
carcinoma. A preliminary report 365

Diets, The influence of certain, upon tumor susceptibility and growth in

albino rats (summary) 100

Duane, William, High frequency x-ray spectra 179

Falk, K. George, Sugiura, Kanematsu, and Noyes, Helen Miller, The influence
upon the growth of transplanted Flexner-Jobling rat carcinoma of
hydrogen ions and of various salts in different concentrations 285

Fat cell, A study of a lipomyxosarcoma with comments upon the origin of
the 109

Flexner-Jobling rat carcinoma, The influence upon the growth of trans-
planted, of hydrogen ions and of various salts in different concentrations. 285

Fowls, negro, Inoculation of sarcomatous tumors into, with special reference
to the significance of chromatophores. 31

Freund-Kaminer reaction, A critical investigation of the 337

Fujii, Tamotsu, and Kon, Yutaka, Inoculation of sarcomatous tumors into
negro fowls, with special reference to the significance of chromatophores . 31

Gade, F. G., Is cancer a biological phenomenon? Some heretic thoughts

on cancer 357

Goldblatt, Harry, An atypical adenoma of the pancreas originating in islet

tissue 277

Greenough, Robert B., and Simmons, Channing C, End-result reports of

cancer operations (summary) 187

Growth of mice, A comparison of the, which ultimately develop carcinoma

with the growth of mice which do not develop carcinoma 7

, of transplanted Flexner-Jobling rat carcinoma, The influence upon the,

of hydrogen ions and of various salts in different concentrations, t 285

, rate of transplanted spontaneous tumors in mice, Effect of a reduction

of lymphocytes on 1

INDEX

381

Heat and radiation, combined, The effect of, upon transplanted animal

tumors (summary) 101

, sensitization to, Further experiments on, by radiation (summary) 193

Heredity, The influence of, in determining tumor metastasis. 139

, The influence of, in determining tumor metastasis (summary) 182

Herly, Louis, A critical investigation of the Freund-Kaminer reaction 337

Heterosexual characters, Relation of, to neoplasms and developmental

errors (summary) 186

Himwich, H. E., Teratomas and their relation to age 261

Holmes, Harriet F., Slye, Maud, and Wells, H. Gideon, Primary spontaneous
squamous cell carcinomas in mice. Studies on the incidence and in-
heritability of spontaneous tumors in mice. Fifteenth communication. . 57

, Slye, Maud, and Wells, H. Gideon, Primary spontaneous tumors in the

kidney and adrenal of mice. Studies on the incidence and inheritability

of spontaneous tumors in mice. Seventeenth communication 305

, Slye, Maud, and Wells, H. Gideon. Primary spontaneous tumors of the

ovary in mice. Studies in the incidence and inheritability of spontaneous

tumors in mice. Fourteenth report (summary) 92

, Slye, Maud, and Wells, H. Gideon, The occurrence of squamous-cell

carcinoma in mice (summary) 180

Huffman, Otto V., Relation of heterosexual characters to neoplasms and

developmental errors (summary) 186

Hydrogen ions, The influence upon the growth of transplanted Flexner-
Jobling rat carcinoma of, and of various salts in different concentrations. 285

Inoculation of sarcomatous tumors into negro fowls, with special reference to
the significance of chromatophores 31

Ions, hydrogen, The influence upon the growth of transplanted Flexner-Job-
ling rat carcinoma of, and of various salts in different concentrations 285

Islet tissue, An atypical adenoma of the pancreas originating in 277

Jacobson, Victor C., A study of a lipomyxosarcoma with comments upon the
origin of the fat cell 109

Kellert, Ellis, Influence of the lymphocyte on the peritoneal implantation

of sarcoma in mice 41

Kidney and adrenal of mice, Primary spontaneous tumors in the 305

Knox, L. C., Massage and metastasis (summary) 192

Kon, Yutaka, andFujii, Tamotsu, Inoculation of sarcomatous tumors into

negro fowls, with special reference to the significance of chromatophores . 31
Krehbiel, O. F., Rohdenburg, G. L., and Bernhard, A., Further investigations
of disturbances of blood sugar equilibrium in their relation to neoplasia. . 223

Kross, Isidor, Cancer and parasite 257

, Effect of blood from immune animals upon transplantable tumors 25

, Parabiosis and tumor growth 121

, Pregnancy and tumor growth 245

Levin, Isaac and Levine, Michael, The role of neoplasia in parasitic diseases
of plants 95

382

INDEX

Levine, Michael, and Levin, Isaac, The role of neoplasia in parasitic

diseases of plants 95

Lipomyxosarcoma, A study of a, with comments upon the origin of the fat

cell 109

Little, C. C, Factors underlying susceptibility to a transplantable tumor in

mice (summary) 106

Loeb, Leo, Further investigations on the origin of tumors in mice. VII.

Tumor age and tumor incidence 197

Lymphocyte, Influence of the, on the peritoneal implantation of sarcoma

in mice 41

Lymphocytes, Effect of a reduction of, on the growth rate of transplanted

spontaneous tumors in mice 1

Massage and metastasis (summary) 192

Metastases, tumor, The influence of heredity in determining 139

Metastasis, Massage and (summary) 192

, tumor, The influence of heredity in determining (summary) 182

Meyer, Willy, A few late results after the radical operation for cancer of the

breast (summary) ■. 88

Mice, A comparison of the growth of, which ultimately develop carcinoma

with the growth of mice which do not develop carcinoma 7

, Effect of a reduction of lymphocytes on the growth rate of transplanted

spontaneous tumors in 1

, Factors underlying susceptibility to a transplantable tumor in

(summary) 106

, Further investigations on the origin of tumors in 197

, Influence of the lymphocyte on the peritoneal implantation of sarcoma

in 41

, Primary spontaneous squamous cell carcinomas in. Studies on the inci-
dence and inheritability of spontaneous tumors in mice. 57

, Primary spontaneous tumors in the kidney and adrenal of 305

, Primary spontaneous tumors of the ovary in (summary) 92

, Studies in the incidence and inheritability of spontaneous tumors in. . . . 139

, Studies in the incidence and inheritability of spontaneous tumors in

(summary) 182

, The occurrence of squamous-cejl carcinoma in (summary) 180

Muscular activity, The relation of, to carcinoma. A preliminary report . . . 365

Negro fowls, Inoculation of sarcomatous tumors into, with special reference
to the significance of chromatophores 31

Neoplasia, Further investigations of disturbances of blood sugar equilibrium

in their relation to 223

, The role of, in parasitic diseases of plants 95

Neoplasms, Relations of heterosexual characters to, and developmental
errors (summary) 186

Noyes, Helen Miller, Sugiura, Kanematsu, and Falk, K. George, The influence
upon the growth of transplanted Flexner-Jobling rat carcinoma of hydro-
gen ions and of various salts in different concentrations 285

INDEX 383

Operations, cancer, End-result reports of (summary) 187

Ovary, Primary spontaneous tumors of the, in mice (summary) 92

Pancreas, An atypical adenoma of the, originating in islet tissue 277

Parabiosis and tumor growth 121

Parasite, Cancer and 257

Parasitic diseases of plants, The role of neoplasia in 95

Peritoneal implantation of sarcoma in mice, Influence of the lymphocyte

on tfee 41

Phenomenon, biological, Is cancer a, Some heretic thoughts on cancer 357

Plants, The role of neoplasia in parasitic diseases of 95

Plasma, whole blood and, The protein content of the, in cancer 127

Pregnancy and tumor growth 245

Prime, Frederick, A phase of tumor biology (summary) 96

, Effect of a reduction of lymphocytes on the growth rate of transplanted

spontaneous, tumors in mice 1

, and Rohdenburg, George L. The effect of combined heat and radiation

upon transplanted animal tumor (summary) 101

, and Wood, F. C, Further studies in radiation dosage (summary) 177

Proceedings of the American Association for Cancer Research, Thirteenth

Annual Meeting 87

Report of the council 87

A few late results after the radical operation for cancer of the breast 88

Primary spontaneous tumors of the ovary in mice. Studies in the incidence
and inheritability of spontaneous tumors in mice. Fourteenth report. . . 92

The role of neoplasia in parasitic diseases of plants 95

A phase of tumor biology 96

The influence of certain diets upon tumor susceptibility and growth in

albino rats 100

The effect of combined heat and radiation upon transplanted animal

tumors 101

Interstitial injections of an active deposit of radium emanation in a rat

carcinoma 104

Factors underlying susceptibility to a transplantable tumor in mice 106

Proceedings of the American Association for Cancer Research, Fourteenth

Annual Meeting 175

Report of the council 175

Report on the Journal of Cancer Research 177

Further studies in radiation dosage 177

High frequency x-ray spectra 179

The occurrence of squamous-cell carcinoma in mice 180

The influence of heredity in determining tumor metastasis. Studies in
the incidence and inheritability of spontaneous tumors in mice. Fif-
teenth report 182

Proofs of the constitutional nature of cancer 185

Relation of heterosexual characters to neoplasms and developmental

errors 186

End-result reports of cancer operations 187

Radium in cancer of the bladder 190

384

INDEX

Proceedings of Fourteenth Annual Meeting — continued

Problems in cancer research 191

Massage and metastasis 192

Further experiments on sensitization to heat by radiation 193

Malignant tumors of the thyroid 194

Experimental production of tumors 196

Protein content, The, of the whole blood and plasma in cancer 127

Radiation, combined heat and, The effect of, upon transplanted animal
tumors (summary) 101

, dosage, Further studies in (summary) 177

, Further experiments on sensitization to heat by (summary) 193

Radium emanation, Interstitial injections of an active deposit of, in a rat
carcinoma (summary) 104

, in cancer of the bladder (summary) 190

Rat carcinoma, Flexner-Jobling, The influence upon the growth of trans-
planted, of hydrogen ions and of various salts in different concentrations . 285

, carcinoma, Interstitial injections of an active deposit of radium emana-
tion in a (summary) 104

Rats, albino, The influence of certain diets upon tumor susceptibility and

growth in (summary) 100

Ray, L. A., and Robertson, T. Brailsford., A comparison of the growth of
mice which ultimately develop carcinoma with the growth of mice which
do not develop carcinoma 8

Reaction, Freund-Kaminer, A critical investigation of the 337

Robertson, T. Brailsford, and Ray, L. A., A comparison of the growth of mice
which ultimately develop carcinoma with the growth of mice which do
not develop carcinoma 7

Rohdenburg, George L., and Prime, Frederick. The effect of combined heat
and radiation upon transplanted animal tumors (summary) 101

Rohdenburg, G. L., Krehbiel, O. F., and Bernhard, A., Further investiga-
tions of disturbances of blood sugar equilibrium in their relation to
neoplasia 223

Salts, The influence upon the growth of transplanted Flexner-Jobling rat

carcinoma of hydrogen ions and of, in various concentrations 285

Sarcoma, Influence of the lymphocyte on the peritoneal implantation of, in

mice 41

Sarcomatous tumors, Inoculation of, into negro fowls, with special reference

to the significance of chromatophores 31

Sensitization to heat by radiation, Further experiments on (summary) 193

Simmons, Channing C., and Greenough, Robert B., End-result reports of

cancer operations (summary) 187

Sivertsen, Ivar, and Dahlstrom, A. W., The relation of muscular activity to

carcinoma. A preliminary report 365

Slye, Maud, The influence of heredity in determining tumor metastases.

Studies in the incidence and inheritability of spontaneous tumors in mice.

Sixteenth report 139

INDEX

385

Slye, Maud, The influence of heredity in determining tumor metastasis.
Studies in the incidence and inheritability of spontaneous tumors in
mice. Fifteenth report (summary) 182

, Holmes, Harriet F., and Wells, H. Gideon. Primary spontaneous

squamous cell carcinomas in mice. Studies on the incidence and in-
heritability of spontaneous tumors in mice. Fifteenth communication. . 57

, Holmes, Harriet F., and Wells, H. Gideon, Primary spontaneous tumors

in the kidney and adrenal of mice. Studies on the incidence and in-
heritability of spontaneous tumors in mice. Seventeenth communication 305

, Holmes, Harriet F., and Wells, H. Gideon. Primary spontaneous

tumors of the ovary in mice. Studies in the incidence and inheritability
of spontaneous tumors in mice. Fourteenth report (summary) 92

, Holmes, Harriet F., and Wells, H. Gideon, The occurrence of squa-
mous cell carcinoma in mice (summary) ISO

Smith, George Gilbert, Radium in cancer of the bladder (summary) 190

Squamous cell carcinoma in mice, The occurrence of (summary) 180

Strong, Wendell M., Is cancer mortality increasing? 251

Sugiura, Kanematsu, Noyes, Helen Miller, and Falk, K. George, The influence
upon the growth of transplanted Flexner-Jobling rat carcinoma of
hydrogen ions and of various salts in different concentrations 285

Sugiura, K., and Benedict, Stanley, R., The influence of certain diets upon
tumor susceptibility and growth in albino rats (summary) 100

Teratomas and their relation to age 261

Theis, RuthC., The protein content of the whole blood and plasma in cancer. 127

Thyroid, Malignant tumors of the (summary) 194

Tissue, islet, An atypical adenoma of the pancreas originating in 277

Tumor age and tumor incidence 197

, biology, A phase of (summary) 96

, growth, Parabiosis and 121

, growth, Pregnancy and 245

, metastases, The influence of heredity in determining 139

, metastasis, The influence of heredity in determining (summary) 182

susceptibility and growth in albino rats, The influence of certain diets

upon (summary) 100

, transplantable, Factors underlying susceptibility to a, in mice

(summary) 106

Tumors, Experimental production of (summary) 196

in mice, Further investigations on the origin of 197

, Malignant, of the thyroid (summary) 194

, Primary spontaneous, in the kidney and adrenal of mice 305

, primary spontaneous, of the ovary in mice (summary) 92

, sarcomatous, Inoculation of, into negro fowls with special reference

to the significance of chromatophores 31

, spontaneous, Studies on the incidence and inheritability of,

in mice 57, 139, 305

, spontaneous, Studies in the incidence and inheritability of, in mice

(summary) 182

386

INDEX

Tumors, transplantable, Effect of blood from immune animals upon 25

, transplanted animal, The effect of combined heat and radiation upon

(summary) 101

, transplanted spontaneous, Effect of a reduction of lymphocytes on the

growth rate of, in mice 1

Wells, H. Gideon, Slye, Maud, and Holmes, Harriet F., Primary spontaneous
squamous cell carcinomas in mice. Studies on the incidence and inherita-
bility of spontaneous tumors in mice. Fifteenth communication 57

, Slye, Maud, and Holmes, Harriet F., Primary spontaneous tumors in

the kidney and adrenal of mice. Studies on the incidence and inherita-
bility of spontaneous tumors in mice. Seventeenth communication. . . . 305

, Slye, Maud, and Holmes, Harriet F., Primary spontaneous tumors of

the ovary in mice. Studies in the incidence and inheritability of spon-
taneous tumors in mice. Fourteenth report (summary) 92

, Slye, Maud, and Holmes, Harriet F., The occurrence of squamous-cell

carcinoma in mice (summary) 180

Wilson, Louis B., Malignant tumors of the thyroid (summary) 194

Wood, F. C, Experimental production of tumors (summary) 196

, and Prime, Frederick, Further studies in radiation dosage (summary) . . 177

X-ray, spectra, High frequency

179

Date Due

— : — i

<%li>un*A* feC**'™ KeSf'/Hfi

m ■

H9T

warn

4i»

Wife

tlMi

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Texts

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Journal of cancer research."

See other formats